Will AI destroy the internet and take your job?

AI generators, chatGPT, stable diffusion, midjourney, open ai and google ai what's the future for you and your job?
SUBSCRIBE on youtube: bit.ly/H2CThat
Support: / h2ct
My Cookbook: bit.ly/ARcookbook
Recipe: howtocookthat.net
Merch: www.youtube.co...
How To Cook That Channel: / howtocookthat
Business enquiries only: [e-mail]
Hi I am Ann Reardon, How to Cook That is my youtube channel this week we take a deep dive into the world of AI, open ai, google ai, chat gpt, midjourney, synthesia, runway. What does AI mean for the future of the internet, misinformation and kids learning at school? Is your job at risk?
Subscribe on email: bit.ly/H2CTemail
SUBSCRIBE on youtube: bit.ly/H2CThat
Recipe details: www.howtocookth...
KZbin Channel: / howtocookthat
Instagram: / howtocookthat
Facebook: / howtocookthat
Feel like binge watching? Here's some more of my videos:
DEBUNKING VIDEOS: • Blossom's Fake Video E...
200 YEAR OLD BAKING: • The 200 year OLD chees...
CHOCOLATE TUTORIALS: 🍫bit.ly/chocolat...
DESSERT RECIPES: 🍨 bit.ly/how_to_c...
CAKE DECORATING LESSONS: 🍰 bit.ly/amazing_...
MINIATURE BAKING: • Teeny Weeny Challenge ...
CLEVER OR NEVER: • CLEVER or NEVER? Kitch...
10 BEST recipes in 10 minutes: • Top Ten BEST recipes i...
HOW TO MAKE MACARONS & SNACKS: 🍩bit.ly/macarons...
You can send letters & stuff to:
PO Box 202
Chirnside Park 3116
Australia
Business enquiries only: [e-mail]

[@lilyh4563]

Yay this just made my day ❤❤❤❤

[@sectumsemparium]

I would never question your Intelligence Ma'am, But i am so shocked you have such through understanding and making such a great informative, educative video and kept it simple. I have seen so many tech bros spewing garbage and here you are making such a great Video. Love your work

[@Grandma_Jizzzzzzzard]

@@sectumsemparium the user is AI.

[@melindarivera6598]

​​​@@sectumsemparium In my opinion, she is the definition of a classy, strong woman, the epitome of what a woman should strive to be: intelligent or educated, kind, gentle, compassionate, loving, etc.

[@sectumsemparium]

@@melindarivera6598 I agree, am talking about how she broke my own stereotype of how is she more knowledgeable than some tech bro pushing misinfo on AI saying, art is dead , movies can be made on AI etc

[@lanasinapayen3354]

As an AI researcher who has to try everyday to explain all of ethical issues with the current popular AI models, thank you for starting with the copyright issue, and for explaining the training process without falling in the trap of presenting the models as actually "intelligent" or having "agency" and making choices like we're in terminator. It's people (companies) making those choices and those choices are not inevitable.

[@redwitch95]

Yeah, I'm not in the field of AI but I have a friend who does adjacent work, and it's been really frustrating how often people assume modern AI is identical to the AI you see in scifi.

[@noxteryn]

The way I explain latent diffusion models is to think about it as an advanced form of autocomplete. Imagine Photoshop removing a person from a photograph, and then automatically filling in that area with things from the background. It can calculate, estimate, or predict what to fill in that area with by looking at what's around it, using probabilistic algorithms. A human can do this with great ease. If you look at a photograph of a person standing in front of a brick wall, you can assume that behind that person is just more brick wall. If you see a patch of red brick wall, you can assume the rest of the wall is made of the same red bricks in the same pattern. That's pretty much what stable diffusion does, using the prompts as a guideline on how to use the datasets.

[@ashemedai]

As an IT guy who works and worked with AI/ML folks, this is why I love the term "stochastic parrots" and I do stress to people that current AI is anything *but*. I prefer to use terms like LLM etc to properly differentiate from "true" AI.

[@tabularasa0606]

Also the danger does not come from AI, it comes from people abusing AI.

[@Powertampa]

@@ashemedai Chatgpt is dumb as bricks if you ask it anything that it cannot find an already answered stackoverflow for. Is a great search engine though. I mean it's a language model, expecting it to have an understanding of anything but basic english is stretching it already.

[@robertbergeron2869]

It is staggering to me that the person who is explaining AI in the most practical terms is a cooking instructor. You are amazing.

[@muenstercheese]

Same here. Ann is amazing at describing complex things in neutral, accurate ways. I love her for it.

[@ByHerHand]

*Food Scientist

[@silkwesir1444]

Maybe it's not surprising. What I find staggering again and again is how good ChatGPT is with creating, modifying, mashing recipes, even though it actually should have no idea what it is doing. I use it all the time now. Learned a few new cooking techniques from it already. Another great thing about it is it's non-judgmental. Want to blend together a Japanese and an Italian dish? Go for it! No "But that's not how to do it..."

[@toomuchiridium]

It does make sense when you realise she’s been explaining some comprehensive issues for a while now. But I agree all the same, Ann Reardon is amazing for doing what she does. 😉

[@Mnstrava]

​@@silkwesir1444 I haven't researched this myself yet, but I'm curious how you would go about using chatGPT for creating reliable recipes and learning cooking techniques. Do you have a background in cooking which helps distinguish what would work or not, or do you just run some trial and error at face value? I'd be very impressed if it's at a level where in a newly created recipe all the proportions are correct! Which is why I imagine it being fantastic for finding inspiration and innovation, but you might still have to rely on your own common sense and subject knowledge? I'm just wondering how easy it is atm to use it for this purpose without having to do additional research to figure out if the output is good or not, especially if you're not very familiar with the topic.

[@Ea-Nasir_Copper_Co]

A couple of lawyers in the US are on the verge of being disbarred because the junior lawyer used ChatGPT to write a legal document in which the AI _invented_ a bunch of legal cases in support of the plaintiff. Neither the senior lawyer nor the junior lawyer checked the cases to ensure they were real, but the defendant's lawyers certainly did.

[@kristincox0212]

Your username gave me a good laugh!

[@TheBaldr]

Let's just be clear they are not in trouble for using AI, they are in trouble for using fictitious cases, basically not checking to see if the AI was correct.

[@ArtisanAsteroid]

@@TheBaldr Let's be honest. If you're willing to use AI in the first place for something like that, you had no business being in your field. In order to make things right, it'd take more work than starting from nothing.

[@Eventide215]

@@ArtisanAsteroid It's much like was said in the video, AI can write these things *much* faster than you ever could. That's why people would want to use it. Especially people like lawyers that are on extreme deadlines. Why it wasn't ran through people to check for validity though I have no idea, that's something you'd normally do anyway.

[@SharienGaming]

bwahahahaha that is some comedic justice XD now we just need some company getting sued because they used chatgpt to generate code and it causing some severe security flaw going into a customers production system XD

[@teomda2975]

I really appreciate how well thought out this video is: Anne started with visual, more concrete examples, and slowly moved towards more abstract ones, to help viewers understand these concepts. Wonderful!

[@lisazimmerman5622]

I ran into the AI making stuff up problem in my chemistry class this spring. So many students tried using AI to do their honors projects. The more savvy students did rewrite what AI gave them, and submitted solid projects with accurate sources..Other students just submitted the AI product with no effort to check it's information, sources, etc. It was a glaring difference and easy to identify. Definitely going to add a lesson next school year on using AI effectively & pushing our admin to incorporate AI into our academic integrity policy.

[@imthecoolest50]

It’s literally common sense not to copy paste from the internet, or AI in this case, and submit it as your own work. That’s plagiarism. Are people actually that lazy and stupid?

[@sylviaramsay9180]

Have had a similar experience in courses about health

[@adde9506]

Any student using AI to do their work for them, regardless of whether or not they checked and corrected it, should immediately fail the assignment. You didn't do the work, you don't get the credit, just like the kid that slacks off a group project.

[@lisazimmerman5622]

@@adde9506 I'm not sure I agree with that. We already see people using AI to do their work for them in the workforce. If a student creates a project and uses AI generation to get them started on a scaffold, I don't have a problem with that. What I want to see is good academic rigor, citation of sources, and synthesis of what they are finding/learning. They should be able to stand on the shoulders of what was created before, as long as they aren't claiming others' work for their own.

[@nonpondo_]

If you use AI to write the base paper, you should write another paper writing what you changed about the AI output to make it factual and proper

[@MasterXenex]

"If it doesn't have an answer in its dataset it will just confidently make stuff up" is basically what got me through college and a couple of jobs so I'd argue that's a surprisingly human approach, all things considered 😅

[@HowToCookThat]

😂that comment made me laugh out loud.

[@MasterXenex]

Yay, success! ❤🎉

[@1nekrus]

And it's all candy and butterfly's, until it you realize machine can spew eldritch amount of text, that can take days, weeks or months to confirm if it's true or wall of text made of stupid.

[@OhSoUnicornly]

Yeah, along those lines also when I've asked it to help with stuff for job applications or similar, even when I give it very specific points to make, it just sounds soooo generic! It's a bit too verbose I think. I asked for a 200 word answer and it insisted on giving an introduction AND a conclusion in that short answer.

[@mrharvest]

Sure. But you wouldn't do that if there was a risk of harm to other humans. You would probably say "I'm not sure." An LLM on the other hand doesn't really consider that - they are trained with the implication that "I'm not sure" is a bad answer (it's really never the answer the user wants), so LLMs generally avoid that and instead just string together plausible words. It doesn't mean that the misinformation / data confidence problem cannot be solved, just that the current training approach has a gap.

[@hollierobinson191]

Hands up to Ann for being one of the best content creators out there. The way she tackles all of these real life topics and makes them educational and entertaining is fantastic. I love how much her channel has grown over the years, she’s truly amazing.

[@jaash7981]

I'm an artist and writer, and I definitely think Drew Gooden has made the best evaluation of AI art. Art, whether it's drawing, writing, videomaking, photography etc. draws from the human experience and their personal views to create something unique. AI can't have experiences or tell jokes or have any sort of nuance. They don't make original things, and they're just glorified search engines. So even if you could write whole books and movies with chatgpt, I don't care. The problem with AI are the big corps who are looking at AI like it's their next big moneymaker, and the people who are callous about self-expression and simply think "well my computer can do that now so artists and writers are useless". Edit to bring some more things up: I don't think AI is going to replace regular art, except in a corporate sense. Things look bleak, but there's still incredible works releasing today in all forms that are getting the praise they deserve. Don't support blatant cash grabs from corps, like terrible remakes and reboots (ahem Disney). Show that you don't want unoriginality Second, there is absolutely a difference between AI and commissioning or referencing. The "input" people put into writing prompts for an AI generator is facilitated by other people's work that's been plagiarized, who often do not consent to having their work being fed to AI. If you have a vision for something and don't think you have the skills to see it through, please commission an artist you like the style of. You get your art ethically, and they get paid. Win win for everyone. Referencing is not simply just copying someone else's work and passing it off as something unique (That's just lying about tracing). It's taking technical cues from them, such as perspective, colors, poses, anatomy, shapes etc. and utilizing them in your own work to enhance your original idea. It's also how we have "Draw this in your style" trends on social media.

[@shadowfox009x]

I also see a lot of youtuber pushing AI generated low/mid-content books like drawing books and children's books and promising their viewers that they can make a lot of money this way. And a lot of people are falling for it and rushing off to throw up their AI book on Amazon. Where they then claim that they own the copyright to the content. I've tried AI for translations and the result required extensive editing to be readable. They showed that it was just a machine translating words but not understanding the context of the sentence and so often context is key. Like you said, the machine doesn't get jokes, or word-plays, or any sort of nuance. And they don't get cultural context (important in translations).

[@littelitt]

Idk what callus means

[@stellar783]

​@@littelitt now this is a prime example of who's at risk of being replaced by AI and who's not

[@littelitt]

@@stellar783 that’s what callus means?

[@Showsni]

@@littelitt Callus doesn't mean anything. Do you mean callous? A callous person is hard-hearted, and doesn't feel any sympathy or empathy for other people.

[@ivanvajs6283]

As someone who does applied AI research I can say that Ann, expectadly, did a great job without being too technical in her description. Keep up the great work! 😊

[@travelwell6049]

She does an amazing job of explaining complex information in a simple way

[@arladeleon3806]

I recently watched a news documentary concerning people using some type of system (possibly connected to AI?) to mimic the voice of an acquaintance or relative to scam money from, especially elderly, friends or family. Some lost their life savings. You're right, Ann, not everyone will use these things in honest ways.

[@anarchy7535]

It's such a frequent issue people have, that they simply don't understand what they're using. Like the lawyer recently who used ChatGPT and then cited a bunch of imaginary cases. It's a CHAT bot. It's designed to talk to you and sound convincing. It is not the same as a search engine. You can't go into it expecting the same thing you would from asking a question in a google search. Even on google you shouldn't trust answers without knowing where they come from. It'd be like having a builder go "I don't get why all these nails didn't go in very well, I hit them with a screwdriver and everything!" It's using a tool for a job it's not designed for, then blaming the tool.

[@noxteryn]

ChatGPT is actually incredibly powerful if you know how to use it. I use it everyday to help with my coding. Yes, if you blindly copy what ChatGPT offers, you will write garbage, but when you have very specific and targeted questions, ChatGPT is hundreds of times more powerful than Google. For example, giving error codes to ChatGPT is much more effective than simply Googling them and hoping to find a relevant forum post by someone who had the same problem. The difference can be solving a problem in a few minutes instead of a few hours.

[@E42545]

Artist/builder and hobbyist researcher- this made me literally lol

[@thetableoflegend9814]

I’m a computer science student and I studied AI last semester. You did a good job breaking down a lot of the issues! One thing I will say is that GPT data sets (although I don’t know about GPT4 specifically) have been scrapped from Wikipedia or Reddit and they continue scrapping all the links it can reach from webpage to webpage as well as some other texts if I’m remembering correctly. This has had many bias issues in the past. There is also 2 interesting bias issues with datasets themselves. Most of the data collected in these giant datasets is not labeled, and (in many cases, there are some exceptions) AI need some labeling to begin training. Labeling is time consuming, labor intensive, and sometimes expensive. This has lead to giant datasets where large amounts of it hasn’t actually been checked by a human. These unchecked parts could contain misinformation, disinformation, bias, and it’s all being used to train. Publically labeled datasets though can also be an issue. An AI needs to have example of words to know what to make from that word. This doesn’t just include nice words. I can’t remember the name of the dataset but there was a public picture database with real humans pictures in it that you could look through. Real people’s images labeled “tweaker” showing a random person who maybe does drugs but certainly isn’t doing any in the picture. They also had categories for slurs. Imagine looking through a dataset and seeing your face under a label for a slur. Data is a huge and delicate issue. One facial recognition company got most of its data by scraping websites which specifically had no scraping rules, and for awhile their product was being used by the government. Do people realize that every bit of text and every image they post is being taken and used for AI training? Would they change their behaviors if they knew that? What if they were given the option to consent? Is it ethical not to give the option? How much of your information is already out there in a dataset you never consented to? These are just a fraction of the questions that need to be asked

[@Liusila]

Aren’t all of these questions redundant if people openly share their data with the world up front? I moved from WhatsApp when Zuccergate happened but many of my friends proudly declared they “have nothing to hide” and didn’t care about their data being used. Welp, this is the price to pay and tbh they should have considered that their images or tags or videos or whatever might be used by whoever is creative and smart enough in the future.

[@torbjornkallstrom2316]

I would certainly have had different concerns about uploading my images to the internet if I could have predicted they would be used in massive data sets to train AI. I bet a lot of artists would feel the same. But now the idea seems to be that we've all consented retroactively because our images are already on the internet.

[@davidy22]

AI doesn't always need labelling to begin training. Text generation in particular can operate with just text input. Labelling is required for classification problems, not problems akin to time-series. If you've just done one semester on AI so far a language model of the same architacture as the LLMs is probably too time consuming, but try building a markov chain model for yourself so you have an idea of how language models are actually trained.

[@MossyMozart]

@thetableoflegend9814 - "scrapped" OR scraped? Huge difference, young AI-bot!

[@CoventSynth]

I don't like the "Some jobs will be lost, but others will come" argument. Yes, some will be able to learn these new jobs and work, but for those who lost their job, it may already be too late to learn it, and they will be left behind. In the end, these people care more about numbers than people, and it shows.

[@kvasir8931]

So how many pieces of furniture do you own that was custom made from a carpenter vs machine milled and sold by IKEA? You had no problem with it until now.

[@cassinipanini]

"some of you will die, but it is a sacrifice i am willing to make" - lord farquad 😭

[@etherealclarity]

The problem isn't really with lost jobs that come from advancing technology - the problem is that we live in a society that doesn't take care of its people. We don't have a UBI or universal healthcare. We don't have adequate processes or systems in place to help people transition from one career to another. In an ideal world, jobs being taken over by technology (presuming it does it well, which is a separate argument) is a GOOD thing, because it SHOULD mean more leisure time for more people, and more freedom and opportunities to pursue things that excite and interest us. But we don't live in that world right now, because as a society we value capital over people and "hard working values" over prosperity and happiness.

[@smolbean2110]

Doing research presentations last semester and same day I gave a presentation on the pros and cons of ai that I tried to be nuanced about, someone else gave a presentation about how ai is great and wonderful and we should all become programmers. Even as a comp Sci major it made my blood boil, people shouldn't be forced into one field because it's all that exists for them

[@ShintogaDeathAngel]

@@etherealclarity so what would people do for money and fulfilling other needs if there aren’t any jobs because of AI? Do we go back to a bartering system or something to get things we need? Not that I believe every single job could be automated, but for the sake of argument let’s imagine AI is capable of putting everyone out of a job. Having a sense of purpose is also important to a lot of people, and not everyone gets that from, say, their hobbies or creating a family (if they are capable of that). Also, what effect would AI taking everyone’s jobs have on the relationship between government and society?

[@janebeckman3431]

I first encountered AI chat in the shape of an AI "therapist" in 1977. Having been in tech since those days, I have watched the evolution of many tools, AI included. The main problem lies not with AI but in the beliefs of its users. As a tech writer, I have learned to spot patterns (including being fairly reliable in identifying the native language of various writers writing in English). I can generally spot the patterns in AI text, and can trace its evolution, since I've watched it "grow up.". It's still essentially a probability generator. It has no concept of context. It does not distinguish Joe Smith the physicist from Joe Smith the dead cartoonist, and can mix and match. It literally does not know what "it" refers to. If I switch context, and "it" now refers to a cartoonist's works, it can hallucinate a reference. There is a case where a cybersecurity expert asked what chatGPT knew about him, and it replied with his obituary, and insisted he died in 2017. Even a nonfunctioning URL was furnished as a reference. But since it didn't know that a URL needed to point to collaborating information, it could not show where this information was obtained. But what if you are, say, a government entity who receives information that "you" are dead? You can see where this could go. (BTW, ask it to parse the "you" distinctions in those last two sentences.) Our biggest dangers are human: the tendency to believe it "thinks" like a human and the tendency to believe its outputs without question.

[@ghostderazgriz]

Not the main issue in my eyes. While I agree that many issues with A.I stems from uneducated use, I believe there are far more significant threats it poses. Even if everyone on earth knew what A.I is and its flaws, could we all agree on how to use it? The most ethical applications that support people rather than profit? Could we agree on sanctions and bans where we feel A.I shouldn't be allowed to roam freely? Could we stop A.I from being used as a weapon for political or monetary gain? Could we detect A.I false outputs before they have the power to cause damage? Even with complete understanding, bad actors exist, and some in very high positions of power. And that's with complete understanding, coupled with what you said, A.I has the power to drastically degrade the collective intelligence of an entire population as more and more people de-educate themselves with its use. Then there's the problem of stagnation. When A.I becomes so prevalent that the majority of data sets on the internet are A.I generated. This leads to endless reiteration of uninspired works and misinformation via confirmation bias. A.I becomes cheaper than paid labor and this stagnates our own ability to inform and entertain ourselves. A.I becomes a necessity to compete in this landscape and originality becomes a scarce commodity with a short shelflife as it's consumed quickly by the unrestricted training modules. A.I is not some terminator threat to destroy humanity true, but it doesn't have to be. A.I can simply be an accelerant. The tool used to improve humanity's ability to harm itself at an alarming pace if left unchecked. It's not there yet, but it's growing very fast and governments already can't keep up. It will only get worse without proper intervention.

[@sleepiestmoth]

THANK YOU ANN!! As an artist the excitement around AI bothers me to no end, because it ignores the concerns about copyright in an industry that takes usage rights extremely seriously (to say nothing of the misinformation and mistakes even a beginner artist could avoid). I appreciate you bringing attention to this, more people need to understand why AI has people from artists to computer scientists raising the alarm, and why it isn't paranoid "we're creating the terminator" panic.

[@group555_]

What issue with copyright is there? Forgery is already illegal. Doesn't matter if you use a quil and ink or ai. If ai is responsible for the copyright breaks if it users. Then so it is the inventor of the pencil. It would mean every murderer is free to go, because it would be the weapon that's guilty

[@mwater_moon2865]

@@group555_ 1. forgery =/= copyright infringement, that's some one selling a piece claimed to be created by some one else. CRI is pretty much the opposite, when you claim work of someone else is your own original. 2. The problem is AI is treated as a black box that's propitiatory info so we DON'T know what it's doing (and neither do it's "creators" because it's designed to re-write itself.) That lack of transparency makes CRI a very hard thing to prove in court, whereas if an artist copies a work, you can prove they were familiar with it and it's more than just derivative (legal threshold in the US is 70%, ie. if I take a copyrighted picture and rip it up and put it back together, call it collage and leave out a quarter of it, it's still copyrighted.) Think cars. If a car designer messes up and the gas pedal gets trapped under the floor mat and people die in an accident caused by the pedal getting trapped, the car company (Toyota in this RL case) has to recall affected vehicles, study what happened, say what went wrong, and develop a fix for it-- in this case they trimmed the accelerator pedals and made it clear that after market mats had to be secured. But when Tesla cars don't recognize flashing lights on stopped emergency vehicles means to be careful and plows into them at full speed and kills people, Tesla pleads "AI" and "it's proprietary" and maybe they will push an update or maybe they don't because it keeps happening anyway and people keep dying. And NOW there's a huge legal battle that's taken up time to force the company that's responsible for putting 2 ton chunks of metal zooming around at 70 MPH in the middle of our cities while telling the owners that it can do all the thinking because it has more sensors than a human has eyes... And lest you put it all on drivers, remember that consumers have the expectation that companies tell them the truth when advertising. One of my daughter's friends is blind in one eye due to cancer, their parents bought them a Tesla so they could "drive safely" but meanwhile a right blind 17 y.o. figures it's all good because the car can see when they can't.

[@sd-ch2cq]

Some of the panic is just 'scare hype' though and should be ignored (an attempt by the AI companies to paint their creations as way more capable than they actually are). The real issues are with things like copyright and racism, and not with 'the algorithm turning self-aware'.

[@nocomment4804]

What Copyright? It's perfectly legal to imitate an artist's style.

[@xr6lad]

@@sd-ch2cqracism? Yawn. Yea ok.

[@rosewilkinson4948]

Ever since your video on fractal wood burning popped up on my feed I've been so enthralled with your videos. Thank you for fitting educational content around charming baking videos, it really is an amazing combo 😍

[@tubthungusbychumbungus]

Its so unnerving going on artstation knowing that by default every post is used to train ai. Because theyre using an opt out system for every individual post, any old posts that haven't been tagged are able to be added to datasets.

[@der.Schtefan]

I'm impressed by your broad knowledge, and am glad that you use your platform not only for delicious chocolate desserts, but also to debunk 5-minute crafts and a broad spectrum of other things on the internet, instead of just "going for the click".

[@Pinkyyyy13]

As an artist, AI is incredibly concerning to me since we are actively seeing companies go to AI for artwork in things like animation and book covers rather than hiring or commissioning real artists. art and artists are already so undervalued and underappreciated as it is, this just makes it so much worse. It's especially mind boggling to me since those AIs couldn't even exist without the real artist that they stole the works from.

[@HeleneLogan]

This is exactly my concern, too. You’re absolutely correct.

[@razmiddle9410]

Well yes. And you're part of the problem if you've ever purchased something made by machines rather than made by hand by an artisan -- a decoration, a rug, furniture, a book that used to be illustrated by hand. This is another shift, just like how textile workers watched the loom take their jobs -- the loom which was built because the creators watched textile workers do their jobs, just like AI being trained on human-created visual art.

[@stargirl7646]

Right! Artists have almost always been incredibly undervalued in society, and this is just adding injury to insult

[@Eventide215]

It's almost like you didn't even bother to watch the video. You saw the title and just jumped to the comments. Also, what you say is pure conjecture and therefore misinformation until you've proven it. Anne actually proved you wrong in the video for why companies don't want to go to AI for artwork and such.

[@ko2vo]

​@@Eventide215 Ann also covered the comic book that was initially copyrighted, then revoked once they discovered it was ai generated. If people are all honest, then sure there wouldn't be any fear, but people lie, companies lie, and there is a monetary incentive to lie. It's naive to ignore that aspect (or you just wanted to be mean to a stranger on the internet, in which case you should be ashamed of yourself)

[@Kyrieru]

One thing to note about Stable Diffusion ai is that it doesn't really learn style, it just learns tags. Ai is incapable of producing poses or compositions which do not exist in its training, and tags are too broad to adequately separate elements from each other. Its why using a tag like "short hair" will suddenly change the entire composition despite pose and composition having nothing to do with hair being short. What you're doing is increasing the influence of the "images of short hair" and all training data close enough to it.

[@Novur]

The most frustrating part of all of this is that all this AI futurism is often dressed up as Star Trek Idealism, wherein everyone can ask a computer to do anything they like, and it just does it perfectly. The reason it's so frustrating is that in Star Trek, nobody has any money, because sentient labour isn't necessary anymore! There are no poor people, because EVERYONE has the magic computers! But in the REAL world, all these hyper-rich tech magnates want to use AI so that they can STOP PAYING PEOPLE FOR LABOUR, and then SELL the fruits of their human-supplanting-AI-labour in order to take MORE money from the general population to make themselves richer. It's just greed!

[@stargirl7646]

Oh gosh, EXACTLY.

[@tcunero]

At some point if no one has jobs then no one has money, then rich people have no one to sell to. The star trek world is a post scarcity one, universal income may indeed need to become a thing when no one has a viable job because all the work is done by bots.

[@littlestbroccoli]

Yes it is.

[@ruthspanos2532]

Yes! It’s all very well to imagine a post scarcity future, but right now our tech billionaires seem perfectly happy to twiddle their dials in ways that increase their profits at the expense of everyone else. We need to wrest the control out of their greedy fingers, so these tools are used in ways that benefit all of us. If these dudes are concerned about AI destroying the world, they just need to limit its scope. Don’t hook it up to real world abilities. Put humans in between before any decisions are made that could end us all.

[@mintish_]

@@tcunero and how many people would starve before that happened?

[@Peaslepuff]

My main concern with AI models, particularly the language models, is not an issue with AI at all, but with a lack of critical thinking in general society. AI, ofc, isn't the root cause for any of these issues - misinformation, copyright infringement, plagiarism, bias, etc. It's just unfortunate that we keep trying to address these issues on a surface level. Which is why I love your channel, Ann!! It is so vital to have people like you - and Legal Eagle, Doctor Mike, etc. - who are out there educating people, asking important questions, making us consider things we might not have realized.

[@arturoaguilar6002]

It doesn't help when the owner of the specific AI service saw pretty much impossible to credit all the people who made the images they were about to feed to the AI, so they feed them without crediting anyone. Can't really blame general society when the key people behind the AI are the ones lacking critical thinking.

[@cassinipanini]

@@arturoaguilar6002 i would say thats not a lack of critical thinking, its actually very smart given the situation. sadly its also unethical af

[@Lord_Slug]

over the past few months I've noticed AI waste in Google image results - if I try to look up a specific but kind of obscure term (one recent example was "realistic alligator in a suit"), over half of the results are incomprehensible AI generated mush. AI generated content clogging up all corners of the internet is already becoming an issue that I imagine is going to continue to spiral as it's popularity increases and millions of gigabytes of it continues to be generated

[@ghoulchan7525]

same with Deviantart, artstation, any art site really unless they have a strict No Ai image policy. is flooded with these things.

[@noxteryn]

I just tested your claim on Google, both with and without quotation marks, and couldn't replicate it. I see a few stock photos, some illustrations, and the rest are just plain alligators. If KZbin allowed it, I would post a screenshot here to prove it to you, but I can't. I can assure you I see no "AI generated mush". My first assumption is that you have enabled Google's personal recommendations based on history, so your Google searches are showing you AI-generated images because that's what you've been clicking on. You can disable this option, and your searches will be fine.

[@Lord_Slug]

@@noxteryn I don't click on them and I don't seek them. to check I went thru my history and my exact search was "realistic anthropomorphic alligator" and it remains true that most results are ai nightmares. feel free to test that query and see if it's true for you. this wasn't the only search result where this has happened but I can't remember what I was searching for outside of that one. see if your results reflect what I'm getting now with that exact prompt

[@noxteryn]

@@Lord_Slug You're right. The query "realistic anthropomorphic alligator" does produce a lot of results from AI-generated artworks. I even tested with with adding negative keywords like "-ai" and still got a lot of AI art. I noticed that removing the "realistic" part helped a lot, and shows cartoony drawings. I will admit that is indeed a problem, though for the sake of argument I would say that the query itself is somewhat obscure. I mean, even without the AI thing, I would have naturally assumed that such query will show furry porn or something. But I will concede the point. Yeah, this is troublesome.

[@stygianoatman]

What worries me most about AI isn't the job replacement, singularity, or copywrite issues. I think the worst part about it is that AI threatens to replace CREATIVITY itself. People can easily fall for the convenience it offers. Why learn to draw when an AI can do it for you? Why learn to write, when an AI can do it for you? To create art is a journey, it's a wonderful feeling to make things. Getting good at an instrument. Finishing a thoughtful essay or script. Cooking a delicious meal for your family. The quality of that music, writing, cooking, etc. isn't all that AI cannot replicate. It cannot give you a feeling of TRUE satisfaction. While the outcome of art is fantastic, and I do love a good song, book, meal, etc. those are all just products, and I don't believe we should treat art (or care about it) for what it yields, but the joy it brings to produce. We are human, we were meant to express ourselves, having evolved to communicate using complex language, and making art on cave walls. AI is a quick and easy way to get the product, without the journey and self-expression. In my opinion, you miss out on a big part of life itself, by letting an AI create something for you. That brain is a muscle, use it or lose it. Lastly, AI has no love for anything, even if it became advanced to mimic it, there's no hormones responsible for love, a machine cannot feel. If a Grandma bakes you cookies, you know she does it with LOVE. If a robot made you cookies, there is no love behind that. Same goes for making AI create artwork of your pet. An AI feels nothing when it looks at an image of a kitten, puppy, etc... how are people okay with this? The use of AI can be problematic and insulting to others, but I think it's equally insulting to the self (if not more so). As humans, we should remember we are alive. We DO feel. An AI cannot make something because it loves you, it cannot be your friend, you deprive yourself of true connection by using it. Creation is an expression of love, I believe. A selfless expression (generally). We sing to express ourselves, and it brings people together because we resonate with those feelings. We cook for those we care about. We procreate because we see the beauty of life, and wish for it to thrive and persevere, like it always does.

[@Stervelar]

Adding to that, people keep saying that Ai take references "just like humans", but I wonder if they actually have ever used references or tried to paint (digitally or traditionally). You can see a picture of a banana, and to you, even if you don't notice for yourself, it's more than pixels and color changes. You are likely to know that bananas are sweet, soft, easy to peel, have a distinctive smell, and are food, all that information goes into understanding the fruit and of course, the way you will depict it (if you don't know about it, you might feel the need to try and know as much as you can), but Ais can't internalize any of that. The visual information of a picture, is just one layer of it's significance and that is lost in current Ai models and I honestly don't spot a "prompter" who is concerned about this lack of versatility. Every style, has an origin, an intention, something that the artist want's to focus on, or exaggerate, all of it has a reason and to reduce it all to a prompt and patterns the prompter often don't even know/care about, is kinda lame to me. When you study another artist, you do more than trying to copy and learn from their work.

[@stargirl7646]

YES!! This is a huge point for me. It has no soul - something we already need MORE of in this modern world, not less

[@stygianoatman]

@@Stervelar Exactly! And I think a lot of artists wouldn't even mind their art being referenced, by a human. Because unlike an AI, the human decided to use that art because they built an appreciation for it. If I wrote a Star Wars fanfic for example, it would be made with the two decades of burning love/passion I have for the series. It's something I actually respect, have reverence for. What I make wouldn't be perfect, maybe not even good, but we all have a heart like the original creator George Lucas. As you said, even a banana is more than a banana to people. Food is *literally* life. Art (the product) is not really life, but the essence of it sure is.

[@stygianoatman]

@@stargirl7646 The saddest part is I've heard few people concerned about it, like frogs in boiling water unaware of the danger (though I know a frog actually jumps out) They can take our money, they can take our jobs, but the worse we can lose is our soul. I think something similar has basically happened with the food industry, fast food specifically. We value a product so much, that sometimes we forget the personal aspect that food offers. It was always supposed to be a more involved and social process. Not that there's anything wrong with a burger, just that cooking for others or even yourself is both a creative outlet, a means of socialization (as is eating the meal), and a way you love yourself. Similar to how a shower is self care.

[@Gamesaucer]

I don't think that the creative process itself is under a lot of threat to be honest. Humans have always needed to keep busy. Leave a human alone in an empty room for a few minutes with a button that gives them an electric shock and they will usually start pressing the button, even though they know what it does. What's more important is that if AI can create images that are "good enough" it might start not just taking job opportunities from artists, but supplanting those artists entirely. On big productions like movies or video games or similar things, it's a corporate process that's prone to executive meddling, which is ultimately going to lead to inferior products just because, while they're e.g. of 30% lower quality, they're also 50% cheaper. And you may very well end up losing that creative spark that makes many of those works so special. It's not really that a human produced it, it's that it was produced with _intent,_ which AI cannot currently hope to do, and likely will not be able to do for decades if not centuries to come. There's nothing inherently inferior about "artificial" life over "real" life... it's just that for now, the "artificial" life we create sucks. With AI taking jobs from us, in the long run it should mean that we become more prosperous. If we have to do less to achieve the same, it means our working hours can decrease without our wages or standard of living decreasing. And if this is true, it'll leave more time to pursue hobbies and rather than worry about what you're going to eat if you don't find a job that pays better soon. I'm just worried that capitalism won't let this happen, as it's built on exploitation of a class-based society, which comes under threat from automation. Either our economy will have to change massively, or capitalism is going to keep from us the most important advantages AI promises in the first place. In conclusion... If circumstances allow, humans will keep being creative. And the more things become automated, circumstances should allow it more, not less. The restricting factor is the way our economy is currently structured, not the increasing prevalence of AI.

[@voodoosleeper]

We can always rely on you for a thoughtful, measured, and well-researched take on any subject you produce a video about. The internet, nay, the world needs more Ann Reardons.

[@WitchbladeEdits]

Honestly what's frustrating me is the fact that AI users started calling themselves "artists". I am following a page on facebook, which posts different AI generated outfits. One day the admin posted a photo of herself and said "the face behind the artist". I can't believe someone that writes a few stuff and lets the computer do the other things, calls themselves an artist, lol.

[@E42545]

This is exactly what people said about the camera, and then the digital camera, and then the iPhone camera lol

[@jaychang6723]

@@E42545 even with camera photos there are still lots of human aspects that go into them: the framing, the backdrop, composition, and you know…physically going somewhere to take the photo. I’d argue comparing all of that to typing keywords onto an AI generator without much modification is a bit of a stretch

[@E42545]

@@jaychang6723 if you saw the way the photography community reacted to the announcement of the first smartphones with excellent cameras, you’d know it’s not. Obviously all art (including photo) requires *some* kind of human input and is subjective even in its definition of art to people’s desire to consume it. Generative Art both requires human input and is clearly something (I’m not saying everyone) people enjoy. This is absolutely the same mindset of all the people who gatekept and screeched “low effort!” and “muh job!” when people started getting good phone cameras and inexpensive DSLRs for Christmas and taking beginner-level portrait sessions for $20. Imo there’s no such thing as saturation unless you’re not special or unique and are easily replaceable to begin with- and life and creating gets a lot more comfortable once you realize that we all are.

[@SnailNick00]

@@E42545 You are comparing photos still taken by humand to pictures generated and stolen from actual artists. Do not downplay this issue you clown.

[@cactustactics]

I'll say there is work that goes into generating these images, and often people are sketching out ~something~ and saying "make this look like my prompt", so there ~can~ be effort and even creativity involved (to a point). But it's the model that's actually generating the image obviously, and are you a "writer" if you give a prompt to ChatGPT and post its results? Are you a "writer" if you curate the results by tweaking the prompt, and only posting the output you like? What if you remove the AI from the equation, and you're just giving a prompt to another human who's producing the results? Are you an artist if you tell someone what to draw? There's an argument that maybe you have a "vision" that you're using someone else to create, but is that really what's happening when you don't have full control over the output, and it's more "see what you get"? What about when people are copy-pasting blobs of keywords in their prompts that other people have worked out give good results instead of bad and meh ones? Personally, what really gets me about people calling themselves artists is when the image has obvious problems, like backgrounds that don't make sense, impossible anatomy etc. If you're an artist, why not take pride in your work and fix it, and only put it out when it's good enough? A lot of it comes off like they don't actually care, they generated something that they think looks cool and they just want to say "check this out!" Which is fine, but it's more like "look what I found" than "look what I made". And sometimes "did you even look at this yourself?"

[@dndndndndn419]

I really appreciate how much detail Anne gave to the source material / copyright issue. Computers are NOT capable of original thoughts or ideas - they require Terrabytes (literal billions of examples) of actual artists’ works to create an average that best mimics a human’s original work. It’s just a chewed up and spit out version of the unlawfully obtained artwork of an actual human.

[@SheyD78]

Trusting information going forward is definitely going to be harder unless serious steps are taken, and just finding things that aren't AI generated is going to be problematic. Kyle Hill brought this up on his science channel, as there are now a rapidly growing number of 'science' channels populated with entirely AI generated videos of questionable value, but put out every 12 hours so as to always have something new to click on. On a different note, what an adorable dog!

[@stargirl7646]

Oh dear…

[@leemasters3592]

That video came to my mind too when Ann mentioned fact checking and ability to trust your sources.

[@littlestbroccoli]

I can honestly see a time soon when it'll be perfectly understandable for someone to say, "Oh, I don't go online" because of the trash heap the web is going to turn into. And that will probably get better eventually, but not before it gets so tiresome that millions of people don't even bother.

[@Eye-Of-The-Beholder]

Oh sweet another Kyle Hill watcher! I think that the fact that we have content creator so different from each other discussing this topic speaks volumes on how sever the issue at hand is

[@stargirl7646]

@@littlestbroccoli hey, I was just thinking about that possibility too! Thanks for putting it into words. I can definitely see a time where there’s so much trash online that we end up going analog again… maybe that wouldn’t be a bad thing lol?? 😅

[@danyakloppers9288]

I really love how you investigate an issue and explain it so simply. You aren't just a baker but a true scientist. Thank you for your amazing videos.

[@erymenn_6541]

Data scientist here. You underline very well the concerns with the creation of AI models. Concerning worries about the AI usage, I want to emphasize AI is a tool. Like a knife, there are OK usages (e.g. cutting veggies) and not OK usages (e.g. murdering people). AI can't act by itself no more that a knife could murder someone by itself. You need people to plug the AI result somewhere for them to have an impact. The people that create the AI are responsible for the data they use and the scope of use they define, while the users of AI are responsible for misuse or out of scope use. This is not yet a legal responsibility, but will soon be (the AI Act). This isn't going to be easy to apply of course, but reminding AI is a tool highlight its usage is a mix of tech and people problem, so the best solutions will most likely be a mix of tech and people solution too.

[@evanmak7837]

Chinese and Korean game companies recently started using AI art instead of hiring human artists. It happened en masse, a ton of people lost their jobs and are now unemployed, unable to find a job again in the industry. The struggle against the graphic artists is more than real. It is happening as we speak.

[@caseygreyson4178]

I don’t want to sound all “holier than thou” but it really saddened me when I was the only person in one of my university classes who actually wrote my final essay. The other 11 people admitted in our class group chat to using ChatGPT (I’m shocked they would admit such a thing without the fear of someone in there reporting them). We all got As, so I feel extremely disappointed and demotivated going forward, if everyone around me is just gonna get the same grade as me when I was the only one putting in hard work.

[@kvasir8931]

Hard work has never mattered. Its always the end result

[@ShintogaDeathAngel]

@@kvasir8931 and sometimes getting a result does actually require hard work?

[@cactustactics]

If it helps, you're at uni (and probably paying for it) to learn, to be guided to develop knowledge and skills and experience. That's the point of an essay, the thought and the research and making your argument. You did that hard work ~for yourself~ and you grew because of it, y'know? Try not to worry about what everyone else is doing, and keep pushing yourself to do the best you can! You'll come out actually having gained something for your money and time besides a certificate Besides there's nothing new here, people paying for essays has always been a thing - it's just now it's free, worse quality with higher likelihood of errors, and more likely to get caught out. And even if they got away with it at the moment, it doesn't mean stuff won't be scanned in future, trip some flags, and have them called in to explain their essay that they didn't write, and what they meant by ~this part~ or why ~this citation~ is made up. This stuff is suddenly kind of a big issue now, so it's gonna get a lot of scrutiny - so it's not worth risking it anyway

[@stargirl7646]

WTF?! I’m so sorry!

[@dondcco]

Results get graded, and you can even be hired on the basis of those results. But you will need to actually perform tasks and make transformative work to actually be successful, which are dependent on skills, which is what you gained by actually doing the work. I hope you don't get discouraged :)

[@yasao_art]

As an artist, I'm very grateful for your video. AI is a BIG problem for the art community, there's no doubt about it - why would people want to pay artists to paint for them, if they can easily generate a picture for free, after all? It's so frustrating and disheartening. News (at least here in Germany) keep ignoring that problem entirely, whenever AI is mentioned, they merely report on subjects such as fake news generated by AI or students cheating, which yes, are serious problems too. But artists are never mentioned and it breaks my heart a little whenever I see AI "artists" with millions of views or likes. I'm a smaller, unknown artist and I've put decades into my craft. Every artwork takes days, weeks, if not months - and here I am, fighting not only against algorithms that hide my art, but against AI too. It feels like fighting windmills and thus I'm often left feeling very, very hopeless.

[@tcunero]

I do understand your frustration. However is there a difference between this and any other job that technology has taken over? A cobbler? Blacksmith? Printing press replacing scribes? Clocks replacing people that used to come to your door to wake you. It just seems that this is that time in history for another skill.

[@MetalheadAndNerd]

Photographers said the same when everyone started carrying a high quality camera in their pocket.

[@mimosa5174]

⁠@@tcunero your view of the world is so bleak!! The examples you cite weren’t exactly how humans expressed themselves. They were jobs that came about because of a necessity and disappeared when the necessity did too. Art and artists have existed since the dawn of humans because we have a natural need/drive to express ourselves so it would be very sad to see them disappear.

[@RejectedInch]

@@MetalheadAndNerd that argument is BS.

[@RejectedInch]

@@tcunero BS. Through and through. Back to primary school, history class will help. Also a class about " how to shut up when not knowing jacksh*t" will help.

[@christinedugas3089]

Great job, Ann! Excellent overview of some of the current issues with AI. I've been in a job that was changed/downgraded due to AI. I did medical transcription, which eventually morphed into "editing" voice-recognition-generated reports (at a lower rate of pay, of course). The voice recognition would "learn" that when a particular doctor in a particular specialty said a particular phrase, what it meant, but it couldn't recognize exceptions. For example, when a doctor dictates the letters E, G, F, and R, he/she almost always means eGFR, which stands for estimated glomerular filtration rate. Once in a blue moon, in a patient with lung cancer, it should be EGFR, which stands for epidermal growth factor receptor. The E being capitalized or not makes a crucial difference. Voice recognition will "learn" that it is significantly more likely referring to eGFR and put that. Just one more example: Doctors frequently say adventitia when they are referring to adventitious sounds heard when listening to breath sounds. Voice recognition software will "learn" through the "editing" process to change adventitia to adventitious sounds. Adventitia, however, does have an actual meaning not related to breath sounds. It is the outer layer of fibrous connective tissue around an organ. An experienced medical transcriptionist can know the difference between eGFR and EGFR and adventitia as slang meaning adventitious sounds and actual adventitia and transcribe accordingly, but AI is nowhere near good enough. Those are only 2 examples from the multitude I remember from my time as a former medical transcriptionist, and yet voice recognition software with its AI component was responsible for a drop in pay such that I eventually left the profession. This will likely happen to more and more jobs as time goes on.

[@joharakiri]

I think the most dystopian aspect of this is that so many people think we should restrict AI just to artificially keep certain jobs viable instead of just distributing wealth more evenly. Because no matter what regulations governments are gonna come up with there will be ways around them and companies willing to exploit gray areas. I reckon the advancements in AI we've been seeing are a huge call for the (frankly way overdue) re thinking of concepts like copyright law and who gets to own what in a production chain. Not only would this benefit workers, copyright law is also the main reason we don't really get to know what's going on "under the hood" with most AI models.

[@stargirl7646]

The topic of AI is seriously becoming a huge anxiety trigger for me - it seems so stupidly impossible to stop “progress” when there’s a profit to be made 😭 I have to hope that more sanctions and guidelines will eventually be added. I have to hope for SOMETHING.

[@rolboo6158]

Same but cuz I'm an artist amd is rlly the only thing I can stand

[@Powertampa]

Try one of the freely available ones. Ask it some hard questions that require more than just a google search and it falls apart rather quickly. They have no thought or actual critical thinking, they just match one side of the equation to the other. You can't make a machine based on TRUE or FALSE logic gates have the nuance of human thought. Don't think that'll happen unless you have a computing revolution that somehow increases the computational power by magnitudes of millions or billions. I'd worry more about having to clean up the mess the current models make and having a boss that's too dumb to know that Chatgpt is a language model not a code generator.

[@krila3978]

​@@Mito-gk8jp Because there's profit to be made.

[@Plum_bird]

Maybe the are just giving us access to faulty AI with the intention of advancing AI to the point it can assimilate more quickly and respond.

[@caralouis]

@@rolboo6158 Im an artist too, i feel you 😭

[@frogbirds]

When I saw this video title I worried for a moment, especially having seen some creators I wouldn't have expected getting really on board with AI and ChatGPT uncritically. Don't know why I worried though, this was a really thorough well-rounded video!

[@muffinsaredabest7349]

Let us all applaud the editing for the introduction lol that was some intense content

[@HowToCookThat]

thanks 😀

[@Pyralis]

Another important thing is that, once a model for art like Midjourney is trained on a specific dataset (Eg, drawings made by Banksy), it is very difficult/near impossible to un--train that model on those images. I say this because some time ago, Deviantart used their art database to train their own AI art generator. People got upset so the site gave users the choice to opt out their images from the AI's training. The thing is, if their images have already been used, then it's impossible to un-use them. It's practically irreversible, like cooking an egg. You'd have to throw that AI model away and make a new one. But companies aren't about to do that and they're definitely not gonna tell thousands of small creators which art was used to train it.

[@Alec_____]

As a writer, thank you so much Ann for bringing awareness to this. It's all fun and games with AI, until real life people and their jobs and livelihood are affected.

[@notsoptimistic]

I'm an artist. ever since the issue of AI taking over the art community with AI it has been bugging me. lots of people seem to hv different opinions on it but for me, it's disrespectful when artists work hard to produce a piece but then AI takes the credit when they're only mixing lots of people art works to make it theirs

[@scw55]

People who use AI to create images are just art directors. They tell the entity what they want and how they want it, and the entity makes it. People who use AI to create images are not artists. For myself, it frustrates me when people post AI-generated images, but the person hasn't gone to the time or effort to fix the image.

[@scw55]

@@michellev2630 I feel like AI art has increased the value of artists posting in progress work and time lapse videos. I don't know if this is on a whole good or bad, but it's changing the content artists post.

[@E42545]

FYI this isn’t what generative image models do- they look at thousands of images and learn what a thing looks like, and then *generate* a completely new image from that information, literally based on a cloud of noise (think like TV static). There is no direct taking or collaging of anything, it’s the exact same way people make art out of “nothing” from things they have seen before.

[@scw55]

@@E42545 absolutely reductive. There is no human experience to filter or stain the work. It is clinical and theft.

[@christopherbennett6571]

@@E42545 Nope. Absolutely wrong. Ai generated images even have watermarks on the generated work to the point where it's even become a default hashtag to say #nowatermark

[@chelseal654]

I’m not afraid of AI, I’m afraid of who is programming it or feeding it.

[@noxteryn]

Most AI projects are open source, meaning that there is full transparency in how they're coded.

[@LeafyK]

Agreed

[@ruthspanos2532]

I am mostly afraid of who is profiting from it.

[@drakewinwest9888]

“I’m not afraid of the super-above human intelligent system capable of coding, complex problem solving and of which I have no idea it’s goals or motivations, I’m afraid of humans but I already understand that threat.” You. An idiot.

[@lbaxel9122]

@@ruthspanos2532 Why? Do you expect work to not generate profit?

[@Alexis-sn5cf]

As an artist who loves to share other creator’s work, it’s becoming increasingly hard to filter out the massive amounts of AI content. It’s frustrating because I don’t believe AI art is art, so being 100% sure of a human hand creating the works I’m sharing is becoming nearly impossible. It’s getting to the point that I’m scrutinising every piece for inaccuracies, which isn’t helpful as humans are capable of errors. I don’t feel good in looking at an image and thinking “Is that anatomy because the artist is still learning, or is it because of AI?”. Just a couple of years ago I could share work with confidence, but now I need to second-guess it all. It’s not helpful for new artists as they feel like their effort isn’t worth it because someone can just type in a paragraph to get something in seconds. I do feel this will negatively effect the art community, and as a result, we’ll see fewer human artists sharing their experiences with the world, which is a really depressing thought.

[@kvasir8931]

Basically "if it walks like a duck, looks like a duck and quacks like a duck Im the one who decides if its a duck or not"

[@wakingcharade]

It's totally reasonable to want to only promote art of a certain type or made by certain people in certain ways, and this has always been an issue sharing art - is this stolen? traced? posted without permission? is the person who made this also drawing terrible things or do they have a history as a person of harmful actions? Do we do research on every image we see or just the ones we share? And what happens if we mess up? The question of what is or isn't art has been debated for as long as humans have had a concept for art. It's also come up every time new technology appeared. Photographers dealt with it to - how can it be art, if all you do is take the image and frame it as art? Having reservations about the ethics of AI generated images, or wanting to uplift art made without it both make a lot of sense. I'm just not sure how, if you cannot tell the difference between two images without knowing the full processes of their creation, you can't say one "is art" and the other "is not" unless you are defining art in some way that requires us to know the full process of its creation. Which, if you are, I'm curious what your definition is -- As I've said, there are hundreds floating around out there at least, so there is not going to be a definitive or scholarly answer on if something that uses AI image generation can be art, I don't think. I also wonder what you think about art that uses AI as a tool, but also involves a human curating and editing and manipulating it in some way? And in which ways make it art (ethics aside) vs which do not? Like if someone asks mid journey to iterate in specific directions and keeps adding new conditions and new curation, that may not be, but if they take the image from mid journey and use it as a reference, that resulting human painting would be. So the answer lies somewhere between them? I'm curious where you think that is. What if an artist feeds exclusively their own art into the program? Or only their own art and public domain works - like from over 100 years ago? I really do hope that young artists don't give up because of this. There are always going to be people who produce art faster and with more precision than you, be that visual art or music or writing. But the purpose of creating art is to realize your own unique vision. I think the flip side to all of this is that if we could make an AI art generator that is ethical - taken only from public domain, say, it could allow more people to realize their creative visions, to tell their stories, to make references, to share ideas, all kinds of things -- that would still not be art made by a human hand, but if it is made to realize a human vision, even if you don't call it art, even if you want to make a space that excludes it from the collection, couldn't it still have value?

[@razmiddle9410]

Visual art will be more of a hobby, like woodworking or rug making. Woodworking is a fun hobby but not particularly lucrative because machines can do it better, faster, and cheaper. Whatever you think is special about visual art used to be seen as a special about many other kinds of creation, there's no moral uniqueness to machines getting involved in visual art as well.

[@lotusdragonjbh]

I feel you. It’s pitiful how many people want to pretend they did something when they didn’t. By definition and the The US copyright court state that these generated images can not be labeled as art , nor copyright protected simply because the were not made by human hands: these programs have no agency or reason and can not explain why certain things were executed like they were. Besides the prompter(commissioner’s) input , which is just a request, no human makes these images. So yes , there is a very good reason to not consider these images worth sharing. People getting hype on these things are usually after the gimmick, not the work itself. And anyone who says these are “better” than actual artists- 1) no they are not, plenty of artists draw better than these things, they do quality over quantity. These programs are spam machines. 2). These algorithms, got this good because the database laundered work from artists that were better. So I agree, why share work completely generated by a computer and have no effort outside of someone having good fast food ordering skills. Only plus is that all these images are public domain. I also don’t understand why ai shills and non-artists are pushing artists to use this technology when they could already draw without it. But whatever, don’t let this tear you down. More people are realizing these negatives and standing up to it. Programs like mist and glaze poison these laundered datasets composed of stolen data, more places are scrutinizing ai and its negative effects, numerous governments are stepping in and there is much talk of regulation, some sites are developing ways to detect ai images and ban it, there is the copyright thing. There is hope. And artists don’t need a glorified pony generator to create, their skills can be transferred to damn near any medium. An ai prompter is useless outside of a computer. Hopefully that is some sort of comfort, sorry for the ramble.

[@bluemoon7785]

I understand your concerns. But shouldn't technology avance though?

[@justbreathe549]

I understand all the logical issues (copyright, bias, jobs, etc) but what keeps me awake sometimes is the feeling that every fuc***g apocaliptic movie started like this.

[@OfficialR3M1X]

The AI artwork problem is massively concerning for artists because artists already struggle a LOT from a multitude of issues outside of AI, and this includes people trying to steal and grift from their work, AI just makes it easier. People have been reselling patreon content or straight up giving out for free for years. Artists have to deal with their works being stolen and traced or recoloured with thieves claiming they have created that work themselves and often need to take action against these people to make them stop - hell I myself had someone steal and lazily draw over an old sketch of mine on Discord, despite the WIP of my image being 2 posts above it, and the thief panicking and just disappearing once confronted about it. The freelance artist market is heavily stacked and still growing, and artists are constantly underselling their work to remotely get their business off the ground (especially when compared to industry standards and salaries.) AI art does nothing but flood an already saturated market with, what in my opinion, is essentially digital forgeries that undervalue the actual work and skill that goes into making art. They solely focus on the outcome and product of the art and never the process or fun that goes into making it. And whilst some AI outcomes can be easy to tell if their made using AI, that argument is not going to hold up for long, and it already isn't. PLENTY of AI users (especially furry related ones) on twitter are mass producing artwork that look nigh indistinguishable from actual pieces. There are already many dramas and communities that are turning on legitimate artists under the suspicion their works may be done by AI when they aren't. And no, learning how to use the programs to produce AI art barely compares to actually learning to draw - using a bunch of code made by someone else, that rearranges noise and pixels from existing assets stolen from the works of others does not make you an "artist." If you know how to mod games like Minecraft or a Fallout game, you can easily figure out how to use Stable Diffusion locally on your computer through process of mucking around and finding out what settings and prompts works best. That is never going to be a comparable substitute for learning all the art fundamentals, years of practice and studying to actually be able to draw, paint, model, animate, etc. Most artists can't even fight for the copyright of their own work like the music industry can because we can't afford to toss around lawsuits everywhere, especially the younger lot in the freelancer crowd. A lot of us are just average people trying to get by, so when we see our work stolen we just have to kinda call it out in public/online and hope they stop out of guilt on their consciousness (which never happens, because they'll usually just grift from another artists work instead or just rebrand under a new account.) And just, us artists don't need "AI art programs" that make stuff for us, because artists literally pride themselves on making things. We love the process, thats why we literally do it lmao. We have been using AI tools for a pretty long time to create shortcuts and make processes easier, such as CSP with the pose or hand scanner that makes the posing of 3D models easier, or interpolation for animators on programs like Cacani, but they never automatically make the things for us because that takes half the fun out of art; MAKING IT. So the people who benefit from AI art generators are literally grifters and commissioners who want something made on the fly

[@stargirl7646]

Exactly! The fun of art is MAKING it. It’s an expression of our human souls

[@TheYargonaut]

By that reasoning neither photography nor collage are legitimate art media. By the reasoning of not being a substitute for skills, digital painting should never have been adopted when acrylics exist, and acrylics should have been rejected when oil paints existed, and oil paints should have been rejected when tempera already existed. Each innovation changes the skills necessary. Let the new tools find their place, and use the tools you find best for your own goals.

[@OfficialR3M1X]

​@@TheYargonaut ... What. The reason photography and digital art didn't wipe out portrait artists or anyone who chose tubes of paint, a brush and a linen canvas over a digital one is because these new mediums opened new pathways for the art world, and *created* new *artistic* skillsets for us to learn and then, use. AI art does. Not. Do. That. It seeks to REPLACE artistic skillsets entirely. The process goes as; Coming up with a series of one to three word descriptive prompts in a string, chucking it into a program, and you get an end result. If it sucks? Try again or use a reference image for the AI to REPLICATE from. Any insinuation that this is some form of elaborate, creative and artistic process is going to insult any photographer, photobasher and painter, and they will go "what in the bum-licking hell are you on about." Art, the general premise, is a human thing born from our psyche. Music was created and evolved because we like when certain sounds are arranged together. Creative writing and literature was born from our desire to want to communicate and portray ideas in a manner that you can make others imagine it in their head, even if it might be different from what you might have envisioned. Visual art was born to portray the things that your mind projects in a way that someone can see it and understand it. And sure, any artist often is going to make a lot based on our existing surroundings and the things we take in throughout our lives. You could say we copy and mimic things a lot when we create, like referencing an image for a drawing, or using an already existing beat or rhythm when creating a new track. And what makes someone 'creative' is to adapt the information we know in a way that has meaning to it's creator, *which is why people pursue to learn how to make it.* AI image generators do not do that because it's a program. An AI program does not have the capability to learn from human experiences, or gain a talent or learn skills, and rather instead creates predictions on other's human experiences. It REPLICATES, not creates. And the person using it? They aren't going to even develop any skills artistically. They aren't learning to draw, or paint, or how they should creatively compose any part of the image because the program just does it for them. Hell they wont even be reaching post 8th grade creative writing levels to create elaborate prompts because all the AI needs is at most a noun and a couple of adjectives to regurgitate an outcome back. Most people I see online don't even go out of their way to think of prompts themselves and literally just ask other people on what prompts they should use! AI image programs aren't tools for artists and it doesn't make people artists. The people that benefit from them are the ones who can't be arsed to actually create something themselves and instead just want the thrill of wanting to have the label from a bunch of people gawking at what they "made" for instant gratification (that probably compares to a fraction of what an average real painter experiences when people enjoy their results and feel that their actual hard work pays off hmmmm.) AI art literally invalidates and steals off of the people who spend years learning and practicing their craft, I've already seen plenty of my favourite online artists have their works funnelled into AI models and data sets, and have their body of work and style REPLICATED down to the very core. Just there, free to download for anyone to use. I wouldn't be surprised if I was a victim of this too in the future. Will I be mad? Ye, for a bit, but I'll just keep at it and evolve my work to gain interest if I have to, and I'll appreciate those who want to experience my thoughts and ideas, whilst someone else goes to become a factory printer who'll just keep making the same thing from a dataset that'll probably get boring to look at eventually. Side note (because I feel like I might get this comment in one way or another;) no this isn't gatekeeping. Anyone has the potential to become an artist and it is easier now than ever with resources like KZbin to teach people, and you don't need much to start out. Grab a post-it note and a pen, or grab your phone camera and snap some pics. Just have the will to do it. You may not like some of your early steps, and it may take some time to really get on your feet, but you'll be able to look back, laugh about it and really see how far you have come, and how far you can go! And lets be real, with the recent CRS report on May 11th says anything, it's that the only place AI art is going to find itself is inside a courtroom for copyright infringement! 🤫

[@luciferschoice]

All of this. Thank you.

[@TheYargonaut]

​@@OfficialR3M1X That's a fine argument that AI art is low-effort, low-skill, or not deserving of prestige, but it has no relation to whether or not it's art or whether ones using the AI are artists by the bare meaning. If you enjoy the other processes, the existence of AI art & artists can't take that away from you. And if other people are happy using it to avoid some of the accidental complexity of other media, good for them as well.

[@Fallinglaughs]

As an artist, something to note about AI is that there are many artists whose lively hoods are already being affected by AI. There are artists who used to do book covers that have lost clients because the client would rather use AI, companies from around the world have laid off all their artists because they replaced them with AI, and there is even an animation that was released in which the background was made by AI. In the credits, it stated that the background was made by "AI + Human" because it required an actual artist to fix it up, yet they only credited them as human, not by name. This is all with the issue of copyright, so it is not really in the future - it is currently happening, which is why so many artists are fighting against AI. The creators of the AI have talked to the US government and there was the suggestion that they should set it so that artists can submit for an automatic Opt-out where it should be an automatic Opt-In (meaning that instead of having to request for your art, writing, etc not be used in the AI dataset that you would instead have to request for your creations to be used in the dataset); Not only that, in many generated images if you look close enough for some of the prompts of "in the style of" you can find that the original artist's watermark is still seen on the AI-generated image, just all messed up since the AI does not actually generate these images so it doesn't really know what to do with the watermark. Another (non-art) related issue is how companies are also already using AI or planning to use AI - Levi's stated a couple of months ago that they want to test AI-generated clothing models as a way to increase diversity with their brand; so instead of hiring more diverse models, they want to supplement them with the AI models. Granted, they said that they didn't want to completely replace the models, but there are still many questions unanswered and many concerns as to what that could open up. Many users of the AI generators have gone after creators with glee stating they were going to lose their job and a bunch of other terrible things. AI as it stands at the moment, has many ethical issues that are of course not only affecting artists but many others. And you did a fantastic job in highlighting a few of the many points!

[@stargirl7646]

AI diverse models?? Omg that’s so messed up!

[@yazdhenab.]

Like I've said in my comment just before reading your, I am an author and I don't have the financials for a real book cover. I can do them myself, but this time, for this series, I wanted something original so I've used Midjourney, paid for the images and go the right to use them as commercial profit (don't worry I'm not a worldwide author, I do not win more than 10 euros per month with my 8 books) and I assume it. Promise, when I'll be rich and famous (what, I can dream, no?) I will use a real graphist for my covers, but right now, it's not even thinkable.

[@katarh]

@@yazdhenab. You didn't need to use AI for that though. You can purchase the rights to a stock photo for distribution, and assemble the cover and text in a free image editing program, without the need to get AI involved. I did that for my book covers when I couldn't afford a painting. That money went to the photographer, where it belonged.

[@KaiyaAquamarine]

Yeah I know deviantart (website many artists are on - for those who don't know) made their own AI system. They announced all artwork on their website currently was automatically opt-in. It took the artists rallying together to get them to change it to an automatic opt-out. But who knows how much of the current art was used...

[@VaultBoy13]

"Not only that, in many generated images if you look close enough for some of the prompts of "in the style of" you can find that the original artist's watermark is still seen on the AI-generated image, just all messed up since the AI does not actually generate these images so it doesn't really know what to do with the watermark." This is a misunderstanding of what the AI is doing. It's not putting together a collage. It just thinks that something like a watermark belongs on the image about in that area from "looking" at other images. Imagine showing a child 100 images and many of them had watermarks, and then telling the child to create something like those images. They're very likely to draw in a watermark. Now, increase the number of images being fed by likely millions of images with a better ability to mimic what they've been presented. The art isn't amalgamation or collage, it's learned. It's just learning the wrong things from the images being used. And, yes, this points to these corporations using copywritten material to train their bots. Using copywritten material to train their bots without paying for it is a big part of the problem. Were artists properly being compensated for their work, they wouldn't have a problem with the algorithms. The problem isn't the AI's ability to generate art. Industrialization and automation have vastly changed the work that people do. They've eliminated many jobs, but have also created others. There's always going to be a space where artists continue to make art and get paid for it. But, there's going to be new spaces and new work that these tools open up. You should be compensated if your art is being used to train AI. But, you should also expect that there's very little that's going to stop this AI freight train.

[@starshine_lue5823]

It's great having Anne always being the calm voice of reason in a sea if craziness. She is always so thorough and sensible

[@ImprovisionFilm]

My first experience with "AI" was a chatbot named SmarterChild. That was in the early 2000s. It was so groundbreaking at the time that my siblings and I had to convince my parents that it was an actual computer and not some weirdo who liked talking to kids. 😂

[@plebisMaximus]

My first experience was with that cleverbot site. Much more A than I, but a really interesting and fun toy at the time.

[@tirsden]

Nothing beat Santabot for old chatbot hilarity. He got shut down at some point way later than I would have expected, but then at some point he started having issues scrambling conversations between different users. He used to sort-of keep track of conversations before that, in the most simplistic way. Probably something in the code would save one set of words and would occasionally spit out, "Are we still talking about [that thing from earlier]?"

[@evergirl1231]

I miss smarterchild

[@mistifeyed]

I used chatGPT once for fun and had it give me an itinerary for a trip to Cozumel and it told me to go scuba diving right after I flew in. That could be very dangerous, even deadly. When I typed back “but you can’t dive within 18 hours of flying” it said “you’re right, it is not advisable to dive after flying” 😱

[@TheBluestflamingos]

Brilliant video, but I feel the point about visual art is a tiny bit understated. Art generative AIs do not think creatively, they combine existing data, so there is no change from the original work they replicate and the finished result. Multiple times, it has been found that Ai has lifted elements from an artist's work, eg a landscape in the background of an epic painting, or a tree with a tire swing in the center of a backyard, and pasted it unedited in the new work. If a human did that, copy pasting a pirate maiden out of a Clyde Cadwell original, slapping a filter on it, and calling it their own original work there would be serious pushback. That's not taking inspiration. That's ctrl-C ctrl-V

[@fennwenn3317]

I'm so glad you mentioned the bias in training data! I bumped into this problem very early on, when I was playing around with image generation AI before I realized that all of them were trained on stolen artwork. I was trying to generate references for some character's faces. It easily generated some for a character who was a white girl with blonde hair. Trying to get it to generate the other two characters, who were a heavyset african american man and a korean person with some acne, was a nightmare. It kept giving me unpleasant caricatures or european features with no in-between. The korean person kept looking like various airbrushed k-pop idols, which was not what I wanted. The white girl took like five minutes and the other two took literal hours.

[@giantduck8097]

As someone who might be replaced by a robot at work in the near future, this AI is starting to scare me

[@xxPenjoxx]

Same. My job is low skill, low income, and my days are spent thinking about how poor I'm gonna be in 5 years' time.

[@Chromaspell]

i dont think a lot of jobs are necessarily going to go away because of robots. half the time the general public can't figure out how to use "self serve" machines, and the other 50% of the time they straight up just break down

[@giantduck8097]

@Eden my workplace is getting four robots to do exactly what I do. Those four robots only need 1 person to keep them running. I was, unfortunately, not that 1 person they decided on.

[@vonsowards1297]

@@danielsolomon3794 🤦‍♂️super cringe take…. 🤦‍♂️

[@OkinoMuse]

Many artist study and learn for years to develop their own style, most of us do not want an ai to learn our style in no time after being fed with our works. I follow an artist with a very specific and cute art style, and someone just to upset her, used her work to teach an ai to draw like her.. People on Etsy sell ai art, making people who don't read every detail believe it is actual art and expect real artist to be done with a super detailed work in 24 hours.

[@mk-aka-morgan8386]

Chat GPT has also been trained with AO3 works, it knows a lot of terms that are only used in those fics.

[@gwyndlin]

Hi Ann - I've got a hereditary hearing loss, and I just wanted to say how much I appreciate the captions on your videos. I'm not sure that regular-hearing people even notice them, but I do, and it makes such a huge difference. Some creators don't take the time to make captions and I'm forced to rely on the automatically generated ones, which are accurate enough but have trouble with non-US accents in particular in my experience (speaking of bias). So, I appreciate that you take the time. ❤ On the topic of AI, can I ask if you're using the Premiere Pro tool to "auto-generate" captions as a starting point and then editing/fine-tuning the results from there? Either way, thank you (or your editors, if you have those?!) for taking the time and for being inclusive of people with hearing loss. It's very thoughtful and much appreciated. Great content as always.

[@shainazion4073]

You do realize that most KZbin channels have close caption available by you just enabling it.

[@gwyndlin]

@@shainazion4073 Yes, I even mentioned auto-generated captions in my comment. They're not perfect.

[@shainazion4073]

@@gwyndlin My Grandmother was totally blind, she went blind in her 50s. I always was aware of others disabilities. I have a deaf cousin also, she taught us some sign language throughout the years. She reads lips though to help when in the hearing world.

[@roastedcats_III]

Hi Ann! I really appreciate the amount of work you are putting in to every single video you make! You have inspired me to become a dietitian and now I’m studying in UNSW- doing a course you introduced to me! Thank you so much🎉 Keep on going!

[@HowToCookThat]

Awesome, well done on getting in to uni to study dietetics. 💕

[@tedioustendencies]

AI is something that I was initially excited for to be getting more maintstream, like when 3D printers became more accessible. I was heartbroken when it so quickly took a turn for the worst. It's being used in such malicious ways, and even when it's used with good intentions, it's not vetted well before being implemented. Hearing about it being used as an eating disorder helpline just to spit out hurtful responses is just one example. AI could be used for so much good, and it should be. It should be used to assist people in their work, their creative processes, and daily lives. But as always, humans are abusing something on a broad scale and regulation is too slow to put a cap on it.

[@RejectedInch]

because it's conceived to be malicious. The old AIs where more search type and where trained on smaller, specific, scale. Those AI were moving within regulations. These "new" AI are built to bypass regulations in order to gather as much data possible in the shortest amount of time possible, and free of any charge, licence and so on. That's how LAION-5b (Midjourney, Stable Diffusion, to name the most known, rely 100% on LAION datasets, just to give you the proprtion of how big is) managed to accumulate 5 BILION images in less than 2 years. Through the internet scraping bots and through thousands of thaousands of "users" that took someone else's work and added it manually to the datasets. Not to mention the medical records obtained bypassing data protection acts and privacy laws simply bypassing consent. AI on medical field has been deployed bypassing all mandatory trials for human sperimentation. These rules have not being broken by accident, but on purpose. Is no coincidence, if you check the resumes of those "AI bros" that they were ALL involved in the cryptoscams, NFT, Web3 etc. They go from a fraudolant scheme to another, faster than the law s can catch up.

[@lilpetz500]

AI and many other technologies are beautiful and have so much potential. But commercialise it and use it for greed and profit, and it speeds up the process of the toxic outcomes that come from that. It reminds me of the industrial era; we made machines to ease the work of production and make it accessible to more people. We can now create greater, sturdier, quicker things with more human time spent on the artful aspects of them...and ended up with sweatshops and modern day slavery, churning out billions of products that weren't actually requested, and spending billions trying to tell us why we do want as many of those products we can get. The only thing we can do to stop AI industrialisation now being able to cheaply automate our thinking and crafting for us, is a massive social movement, as many people as we can get refusing to make the use of AI not profitable as it is currently being used, to force it into a humanitarian role as it's only choice besides obsolescence.

[@S3anyBoy]

3d printing guns is true freedom

[@PippetWhippet]

To be clear, the eating disorder one wasnt an ai powered chatbot. It was just a regular old fashioned chatbot. That doesn’t make it any better, but you can’t blame that one on ai. I think it makes the point though that demonising the technology is a distraction. Anything ai can do is already being done by other means. Ai is just the latest buzz, so it’s getting all the attention, last year it was the meta verse and the year before it was microled. Next year, we will care as much about ai as we do about whether Mark Zuckerberg is going to destroy personal relationships and it will all feel very silly that anyone cared about that specific piece of technology when our workplaces seem to think that there is still a lockdown and forcing us into zoom. Forget the tech and focus on the actual problem.

[@hensonlaura]

​​@@PippetWhippet Pretending there isn't a problem because other problems will come later is naive. The problems with governments and giant companies is that they're made up of people. People are self-interested & most in their position think nothing of manipulating others, controlling what they know, to also control the decisions that they make. In the United States we've had this problem with social media online and corporate news outlets that refused to allow certain information to be published, in order that public opinion wouldn't be changed by what people knew. Dubious or false information has likewise been pushed out to the fore when it was easily proven to be untrue, but these companies ignored that; they wanted to change public opinion, public behavior. This has affected our political scene in an unprecedented way and one reason the average American is so gleeful about Twitter being taken on by Elon Musk - who will not allow censorship. I've had to leave KZbin often & go to other websites because KZbin will censor what is said in content! The company has a political bias that will not permit people to offend it's own agenda. They won't admit this bias, so right there I know they're cool with lying to billions of people. Likewise Facebook. When governments and companies start using AI against the people? We are going to be in the dark, just where they want us. All any of them want from us is to take our money so they can live the lives they want to live. That includes all governments & all of these major corporations. If people would just believe the evidence of their eyes and what is right before them, instead of what they were TOLD. My mother said when I was little, 50 years ago: "Don't listen to what people tell you, WATCH WHAT THEY DO." Try that.

[@sokolov0]

ann you're a cyberpunk hero, fighting against corporations. thank you a lot!

[@HeroicLeigh]

That comment by the Open AI CEO at the end shows a complete lack of understanding of creative jobs. I love writing. Writing is my passion. Being replaced by AI with my only job opportunity this year of ‘training AI’ for less than half what I’m worth in pay is not a ‘better’ job. It’s soulless. It’s depressing. What I love to do, what my passion is, is being stripped away and I’m left having to reinvent myself to start over again doing work I don’t care about and don’t enjoy.

[@serendipityshopnyc]

May I just say I love the way Ann incorporates her family into her videos. right down to her dog.

[@tabitasutharsan_winklemon33]

I really just want to thank Ann for educating me for so long, I started watching you when I was in yr 8 now im halfway through year 11 and I haven't missed any of your uploads since then. thank you ❤❤

[@HowToCookThat]

Thanks for watching for all those years Tabita 😀💕

[@KaiseaWings]

The difference is that AI creates nothing new, and only remixes. Therefore it needs real artists and people in order to exist, and it threatens their ability to create what it needs. It's a parasite. AI could have some really useful and fascinating uses if we do it right. I use character/./ai or Dreamily for creative writing, but it's hard to know where their training material comes from. A lot of it is my own work and it learning from me, but not all of it, and that's the problem.

[@ribeyesteak9641]

but humans are pretty much the same, our brain doesn't develop without influences from around us so you uncounciously take a lot of inspiration from other things people have created

[@MsVersalia]

@@ribeyesteak9641 that is completely different from a computer copying a bunch of pictures. The human mind does not work like a computer and there are many, many small decisions each artist makes every time they make a piece of art which makes them unique and interesting. AI can only produce a mathematical equation that "averages out" existing imagery

[@PippetWhippet]

@@MsVersalia what you said about how ai works isn’t correct. Funnily enough though, say that exact statement to a neurologist about the human brain and they will assume you have a degree in neuroscience, because as reductive as you think it is, it’s exactly how the human brain does work on a neuron level.

[@jasonpatterson8091]

When robots were replacing humans in manufacturing, the people who weren't being replaced confidently told the workers that there would be even more jobs than before because 'someone has to make the robots and keep them working" and that the new jobs would "just require retraining." A couple of decades later and the number of manufacturing jobs has massively decreased in the developed world and the people that were working them are now working in the service industry if they're lucky. Now we're teaching machines to think for us, and the machines that are pretty bad at that job are scaring people enough to ask these questions. Those people who confidently predicted more jobs? They're getting replaced by this stuff in 20 years. Worse, we're actively developing AI weaponry - machines fighting humans with nobody in the loop to decide whether the trigger should be pulled or not. It's like we want Skynet to actually happen.

[@kloassie]

10:30 Exactly! It *_shouldn't_* have been used as a food colour, but nevertheless it *_was._* That's why it caused those deaths! Deaths and tragedies are not caused by using proper ingredients. They're caused by using the *_wrong_* ingredients. You of all people should know that Ann

[@metaknightmare1234]

First of all I'd like to say good video from Ann as per usual. As an artist though, I'd like to give my 2 cents as this has been a major topic in the community. First and foremost, I want to be clear that the AI in the format that it is now is unequivocally bad. The reason for this isn’t because what a lot of people like to argue about in the news or social media or other places. It’s because as Ann touched on in her video, AI as it functions now is entirely built off theft and stolen work. People want to argue a lot about philosophy and what "art" really means and the human value of creating art and etc, but that is besides the main and primary issue that everything "generative AI" creates is 100% built on other artists' work, most often, *stolen* work which they are not compensated for. Proponents of AI try so hard to dress up and hide this fact behind fancy buzzwords like "oh you just don't understand stable diffusion or machine learning or algorithms" or whatever other new thing it supposedly is that makes it not theft. But repeatedly several times, it is obvious to anyone with functioning eyes that supposedly "original" AI generated images are 1 to 1 matches with STOLEN artwork under copyright and nonconsensually taken without permission from the original artist. If you are doubtful please just look up examples on google and take a look at side by side comparisons and it is boggling that people will even try to attempt to argue otherwise. Even giving benefit of the doubt, AI and computers do not do what you “want” them to do. AI and computers do what you *make* them to do, and whether consciously or not, they are 100% absolutely made to steal work from real human beings. Let's be very clear, the AI is not "making" anything. It is fundamentally different from when a human artist takes inspiration from others. When it completely screws up hands with 6 fingers or makes weird flesh blobs of anatomy, why does it do that? Is it because the AI "thinks" that's what hands look like? Is it trying a new expressionist take on body horror and aesthetics? No, it's because it literally doesn’t have the capacity to understand what a hand even is. When it puts gibberish scribbles in the corner of its image, is it because the AI is being cute and wants to “sign” its work? No, it's because it doesn't know what a signature *is* and there are so many stolen pieces used to build it that it even tries to copy the signatures of the artists that it steals from! The best analogy I have for it is imagining going to 2 baker's shops. You take 2 cakes (without paying btw), cut them in half, and then go "look everyone, look at this cake that *I* made" and then reselling the 2 cake halves to other people as an original work. Obviously people are going to look at that and call it nonsense because it's obvious that you just slapped together 2 other people's work. But then what happens if you do it with 4 bakers, and cut the cake into quarters? What about 10? 100? What about a million tiny slices cut and re-pieced into a Frankenstein cake? What if the slices become so small, that you cannot even tell with the human eye anymore? Maybe you, a human being can't cut a cake into a million pieces to splice together, but a computer sure can. And that in short (though it is a bit more complicated than that) is exactly what AI "art" is doing. It does not “create” anything new or original, it is 100% derivative of other artists’ work. And most notably, take into consideration how this business model is 100% unsustainable and impossible without taking other people's work that was already created - it cannot and does not actually "make" anything from scratch. People want to argue about "taking inspiration" and "learning" as if the machine or AI is a person. But the fact of the matter is it's not. It's as simple as that. It does not have the capacity to learn and function like a person or create art like a person or "create" anything on its own, period. It is the fundamental difference of a person going to a bakery, discovering they like the way a baker makes a cake, then attempting to mix batter and bake the ingredients to recreate the cake they like vs the plagiarized Frankenstein cake scenario I explained above. There is a saying in art that “everything is derivative” in the sense that more or less everything in the art world is built from what came before it. Similarly in writing, it’s not a coincidence that many stories contain many shared elements to the point we even have a term for it - tropes. In the musical western 12 tone scale, there are only so many combinations you can put those 12 notes to make a tune, and things like sampling are even welcomed and encouraged. But there is a big difference in using samples in music or taking a general story structure of the hero’s journey, or using tropes like found family or villain redemption arcs vs literally copy pasting other people’s stories and characters into one you call your own using a thesaurus to change a few of the words, or tracing over another person’s artwork and then calling it original, or splicing 8 other musical artists’ songs into 8 pieces then mashing them together and calling it a new song made by you. To make derivative works transformative, you have to actually *do* something to make them different from the original art, that’s why things like remixes and covers can exist. But AI does not “do” or “create” anything. No matter how many cakes you slice and mash together, you don’t actually bake anything. Most notably, if those other bakers did not exist, your entire product, your AI “artwork" also would not exist. It is a shameless bastardization of an industry that solely relies on stealing the work of others in order to function. And this is what separates from things like human art, music, or storytelling. Even without those who came before, even if Shakespeare did not exist, even if Da Vinci had never painted, people who came after can and will still find their own inspiration and create new and interesting work, even without literally taking from other artists’ products. Without actual real artists to steal from, AI has nothing to offer. With human artists, even without access to other work, you can learn, you can create, you can conceptualize and think and understand *what a hand is.* You know that human hands usually have 5 fingers, even without having to consult and get inspiration from other artist's works. And everything all the AI proponents are trying to do is disguise and hide that fact, to treat it as if taking other people's work against their permission and putting it through a machine that hides the slices it steals is the same thing as a human being who can create, and that the obvious theft going on is okay then because of it. I am so tired of the AI “debate” being framed as some deep, nuanced issue about machine personhood and art philosophy and whatnot when in reality we’re nowhere even *close* to approaching that level of machine consciousness or development yet. The fact of the matter is it’s all just exploiting artists and creators and wanting to steal their work without having to pay them. Do not let them distract you from that!

[@keatonwastaken]

For the work replacement part, AI and Automation is not the enemy of the people, most AI fear is misguided, it is big corporations and the systems we live under where you're not seen as worthy if you're not exceptionally productive and don't work yourself to death. Automation shouldn't mean that people will be replaced in the workforce and become poor because of it, automation should mean that people will work less while being paid the same if not more.

[@Ehh97]

Just the term A.I. is starting to become an anxiety trigger for me. Half the time I have to hide videos that even mention A.I. in the thumbnail now (especially the fun ones where they are especially clickbaity ones that say things like "is art DOOMED?" and crap like that" I am still trying not to completely avoid videos on A.I. because I know that if I am completely avoidant it will make my anxiety worse, so I only try to click on videos that aren't trying to trigger me. Thanks for being informative!

[@stargirl7646]

My comment was similar to this! I’m glad I’m not alone - it’s rough 🥺 And yeah I clicked on this one because I knew she would probably have a good balanced take, which she did. BUT now I’m stuck scrolling the comments endlessly which is NOT helping my anxiety lol! Helpppp

[@Nosteponsneksss]

I appreciate how calm and articulate your commentary is This video eased my mind a bit and you provided valuable arguments against the ethics of AI I’m very tired of the hysteria

[@tedscout4304]

Ann! I came back to watch this again tonight - 8 days later. (I always watch your videos within about 3 hours of you posting). As I have noticed a change in what KZbin is suggesting I watch. You are so right about the algorithm and AI - thanks so much for educating me. I started watching because you made amazing cakes and debunked a lot of stuff I got wrong when making cakes. Then I loved your miniature stuff and now I get educated beyond things I would have thought about. You and your family are amazing - you always explain things with common sense and without mumbo jumbo. I will now watch 10+ old movies and lots of cooking shows so KZbin resets my favourites. Keep making content - you are the sane person in a weird world.

[@vjhreeves]

Just a note on AI's artistic copy accuracy...I'm a professional artist & illustrator. It did not seem to understand Eric Carle's specific collage style. Carle paints paper and then cuts up that paper to make his images. That's not what the AI image showed at all.

[@thesouthafricanviking]

As a 'coder' myself, I can tell you Ai is great at helping to solve issues, but you still need a fair understanding of coding, to be able to see if the response is correct, I have gotten countless issues, where I normally get the response of 'Sorry for the confusion. . . ." When I correct the AI or call it out on it's mistakes. But I do see the possibility of AI reaching a level were the responses are correct every time, and that is more worrying that incorrect answers, I believe

[@Plum_bird]

That’s what I was saying. At the rate it corrects / learns these issues will become non existing. I’m not sure what the shoe-sides goal of the creators are but it’s most likely money driven. I’m worried about society on all levels.

[@drakewinwest9888]

Yep, coders who think coding is safe because it’s not full proof now are clueless. We are using the AOL Online of AI on a problem that wasn’t even built specially to code, and it’s ALREADY impressive. I think it’s only a matter of time.

[@katydid5088]

The other interesting question is "Once it learns one type of code, can we invent others?" When it comes to AI, the data set need to be labeled in order for the processing to even begin sorting and correctly attributing what it is working with and how it should be organized. While it gets rid of one type of job the need for others to present, change, or invent new ways to solve a problem from the ground up isn't going to leave. Add to it that error checking, especially in the early development of AI or different things like interpreting treatment recommendations or lab results, would still require a human interface with all the same pre-requisite knowledge you would need beforehand to make the determination. Humans are still a failsafe for whatever we happen to do, and some of them, especially considering the security concerns of an AI deployed recording system ect, means the human understanding of the subject at hand is still essential. The creative parts of AI, as Anne mentioned, aren't self-directed yet. Decision trees in AI basically give a "Yes or No" answer that then goes through a ranking hierarchy based on perimeters that AI sets up for itself with a human controller. The question of creating Alan Turing's "Thinking Machine" is a ways off (if the last technological doubling is anything to go by, in the next 30-40 years). This is to say that AI and Algorithmic models can't work without input. Therefore, creating a limit to data information in A.I or a more secured world where we still rely on paper inputs and other "non-tech" solutions to create something, is still remains part of our future. I would posit, than a world strictly ruled by Terminator esque robots.

[@jonas8708]

I'm not super worried about that particular scenario. Even if it's correct 99% of the time (which is still way off, it can barely understand basic instructions if it doesn't have literal examples to go by,) software engineers still have to build complex solutions, talk to clients, each other, management etc. and try to balance all the needs for the product they're working on. I mean sure, if all you do is center divs all day or write algorithms thousands of other people have already written, then yeah, you'll be replaced, but I don't think anyone actually does that. E: This all goes for artists as well. Art isn't just "given input A, generate output B" and that's all these models can do

[@Cecona]

The issue I had was how some AI would take pictures from artist and manipulate them into a new image rather than learning from the images. I’m not sure if that is how it still works or if the way AI makes images now is different. Either way it’s scary to think my work as an artist could be so easily replaced by AI. It’s already an issue with people not wanting to pay artists what they’re worth.

[@MsVersalia]

That's all AI can ever do, it's making a mathematical equation out of a bunch of images that already exist. Without people, the AI wouldn't have anything to copy

[@cosmicsoulsorcerer]

Isn't ai art already "inbreeding" with itself?

[@SurpriseKidsFun]

Amazing job as always Mum. Your research & editing was epic on this one

[@becp488]

Aw, I wish my kids had a mum as great as yours 😂

[@iamwarlock1292]

This very morning, my optometrist and I had a similar conversation. His AI software is incredible but he said, in the end, he relies on his own judgment as it is TOO analytical at times and does not pick up nuances that a human automatically would. Great video!

[@Amatureb]

The timing on this was seriously perfect, as in my debating club yesterday this is exact what we were talking about. Quite an interesting discussion.

[@Graylord88]

I've noticed more and more people using AI to sell fake art, and the mom and pops running those cutesy local dog groups on Facebook etc don't have the experience to look out for it or the ability to recognize it.

[@Vickie-Bligh]

As with everything you do, this was presented succinctly, clearly, and intelligently. I enjoy your channel so much, Ann. Thanks for this.

[@FallenAngelHiroko]

AI is a double-edged sword. Yes it can be beneficial in the right hands. But we also know it can be devastating in the wrong hands. And we all know there's going to be "that guy".

[@FerintoshFarmsPhotography]

Thanks for not going all SKYNET with it like so many have lately, you're so smart.

[@briansarahgillespie4299]

You know, for a food scientist, you are equally a scientist of many other topics. Thanks for sharing your thorough research! I learned so much!

[@sarahm2005]

In the world art, I have noticed that the people who are excited about ai are the ones who can't draw two straight lines and the ones who are worried are the ones who have spent years of their lives working on their skills

[@hollyhartwick3832]

There's a bigger problem than just imitating styles. There's been blatant copyright infringement produced by these AIs. A lawsuit was filed over a book cover that was "AI generated" that was barely any different than the original work it sourced and it was obvious it was using someone's art (who didn't submit their work for such use) as a template. Also, Google AI is an idiot. I watch atheist and skeptic channels that debunk religious fallacies. The algorithm occasionally recommends religious videos, and plays ads for religious institutions and Christian colleges. It's really obnoxious.

[@stargirl7646]

Yeah I was shocked that apparently the very dumb KZbin algorithm is AI 😅😅

[@rinlila6974]

Maybe it’s trying to give you opposite views to avoid an echo chamber

[@BewareTheLilyOfTheValley]

I watch anti-MLM videos and the algorithm puts MLM ads on those videos, because it can't tell the difference in the audience it's trying to target. So, MLM companies are wasting their money by having their ads shown on videos for an audience that's not likely to get into them and we love it 😂

[@elainelouve]

@@rinlila6974 that requires the kind of intelligence a robot doesn't have. Also why would advertisers want to waste their money on people who aren't in their target group? I'm a happily married atheist, but still get ads for a christian dating app. Though it can be argued that the AI doesn't know my marital status, but still, it's just a waste of money. Companies say they're trying to offer very targeted ads (because they bill the advertisers for views), but for the moment they're often failing. Btw they're also advertising dog and cat food for me, though I haven't had pets for over 20 years. That's besides the point, but just goes to show the AI isn't as powerful and allknowing as many people claim.

[@cassinipanini]

copying to the point of even replicating of signatures

[@animal_cookie]

(Materials) Chemist here! I was just hoping Ann would do an updated AI video and my wish came true! From a technical standpoint, I'm not too concerned about the current state of AI/ChatGPT replacing the workforce. I've seen so much output that sounds really well-worded and convincing IF you don't already know about it. If you do know, you're quick to see it's both error-filled and lacking of an substance (lots of fluff and filler). But if you do already know, it's easier to just do it yourself rather than painstakingly check ChatGPT's output (in my opinion) From a practical standpoint, though, I only keep that level of scrutiny in my work. I don't think it's possible to analyze every little piece of information we get every day for accuracy. I know I've been known to have an emotional reaction to tweets that were later shown to be fakes. And it takes only a little bit for misinformation to spread like wildfire, and almost impossible to correct.

[@LukaLupinBlack]

Thank you so much for talking about this! I write fanfiction and publish it to ao3 (which is accesible for everyone, even if you do not have an account) and the fanfiction community is pretty much sure that our works were used to train AI's like ChatGPT. People tested this by giving a very specific prompt that was basically the discription of a fanfiction or a specific scene in fanfiction and depending on how many works exist that are similar, sometimes the AI would just say their exact words back at them. Hearing that left me feeling so disgusting, like I'd been used. It felt really good to hear you talking about the issue

[@ijlayugan4149]

Thank you so much for this Anne. When I was a kid I used to think AI was cool. Im now all grown up and a passionate artist, AI is now a threat to us. Not just artists but voice actors, musicians, photographers- just a widespread of mediums! You see these people winning contests from generated content and its just disgusting. Now its AI bros harassing creators being "adapt and let it go" for heavens sake it makes me sick.

[@EngineerOfChaos]

The story you were talking about around 13:30 is something I knew for a long time. My entire Master's Thesis was all about explaining how biases in your training dataset could screw up machine learning algorithms. I was doing speaker verification and the training data was entirely based off of English speakers. And I was able to show just how poorly it was able to understand someone talking another language. In fact, if the person they were trying to identify wasn't speaking English and the speech sample it was being compared again was speaking a different non-English language it literally never got that right. I published my Master's Thesis back in 2012.

[@dreamergirl247]

It has caused a stir in fan spaces as well. Archive Of Our Own, one of the most popular fan content hosting websites, has confirmed that their website has unwillingly been used to train AI. It means thousands, if not millions of fanworks were fed into the machine without any consent from the original authors. And while we’re talking about people using AI recklessly, some people have fed unfinished works into ChatGPT so that they could ‘finish it’, once again, WITHOUT consent from the authors. If we’re getting the short end of the stick as fan creators, I can’t even begin to imagine what it means for people who do this for a living. Scary to think about.

[@megatrongriffin3781]

Hi Ann! I just wanna say I've been with you since your very first video and you don't know how much you and your family brighten my days and illuminate my nights. You guys are shining beacons of happiness, the model family. The amount of success i wish for you is immeasurable and I do have both the Deep Enders and Crazy Sweet Creations, hoping that some of your happiness and glory will rub off on me lol. I love you Ann. Your voice calms my brain when it screams for tranquility. Signed a 23 year old from Jamaica. My name is Ashne McLeod. If you think you haven't inspired ANYONE during both of your journeys, you are sorely mistaken. I love you guys, don't ever stop shining. :)

[@mikerichards6065]

AI might be good at some things, but it will be a long time before it can be an empathic and engaging communicator - Ann, your job is safe. As for these large language models - they are a long way from being general artificial intelligences. They have no deep understanding of what they are producing, they are essentially following statistical weights between the words in their training sets to produce the words that are most likely to follow in a sequence - not the word that is most meaningful, or even correct. They can do a lot, but they aren't smart.

[@noga9895]

Already in the publishing world, I've seen graphic designers and artists be pushed to the side in favor of AI-generated covers, both in indies and even large, traditional publishers like Tor. A lot of creatives are fearing for their livelihoods right now.

[@nautil_us]

Thanks for the comprehensive AI overview! It's both great for people with little previous knowledge, and for people who understand how the technology works but don't know all ethical consequences.

[@Allycat101010]

You may notice that MidJourney in particular seems to do it's very best work when asked to create a forgery of a particular artist- that's because that is fully the intent. It's almost unquestionably a forgery program, that can also do some more generic stuff if you don't ask it to imitate any particular artist (although that will likely still look like someone else's style- it is essentially incapable of 'making up' a new style of it's "own" it's such a dedicated piracy bot).

[@Allycat101010]

And yes, it's fine to be 'inspired' by a style... but forgers have historically made "new" vermeers and renoirs to 'discover' and then sell, almost more often than copying a work so they can replace the original with the copy. I have like literally absolutely no doubt that that's the goal here. Wait a few years, there'll be news about art by a famous great discovered to actually be an AI forgery lol. Bet you anything.

[@PippetWhippet]

You can make a better forgery of digital art with the copy and paste function. Unless you think that making an art piece in a similar style as another artist is “forging their work” in which case, I’m sorry but you’re calling basically every artist since time began a forger. Just trace the history of rock music, from Robert Johnson through to Muddy Waters. Now step through to Johnny Cash, chuck berry and the shadows. John Lee hooker, the animals, the kinks. Dick dale. This is just a tiny time period but you can see rock music develop and grow. If they didn’t “forge” from each other, we wouldn’t have rock music. No-one develops their own style in a vacuum, people copy that which inspires them. And from that, settle into their own niche.

[@Gamesaucer]

There's an image model called Mitsua Diffusion that's been trained on only public domain / volunteered images. It's leagues behind what things like Midjourney can produce, of course, but it does mean that there are people who are beginning to think about how to create these AIs in an ethical way. That does not address the full range of issues with the technology, but it does address most of the concerns from artists. I wonder if we might start seeing similar models in the future that disclose all of their training data and have gone to great lengths to ensure that it is free of copyright or licensing issues. As an aside, I definitely appreciate the down-to-earth approach here, rather than fearmongering about AI's capabilities. As it stands, AIs are rudimentary at best, and can only excel at very specific things. The problem isn't that AI is smart, it's that humans don't know how to deal with it. And with AI-generated nonsense potentially polluting future AI training data, it might spell trouble for the capabilities of future AI. But we'll have to see where that goes. For the time being, humanity's most dangerous threat is still humanity itself.

[@emilyhunter666]

I'm a casual writer, mostly doing fan fictions. Every time I prepare to upload a new chapter I'm anxious about what AI could do. Sure, I've set my writing page to only be visible to registered users and deleted anything that might be online elsewhere, but I've also come to accept that the only way I can be certain that my works can't be scraped to train AI is to delete them all from the internet. I don't want to do that, I know I have international readers and I would hate to suddenly disappear from their lives.

[@Queezbo]

There are few things that set my heart at ease more than a calm, thoughtful person saying intelligent, well-thought, and reasonable things about issues that matter. Needless to say, I love Ann Reardon!

[@Randompuffin1]

My mom was a journalist and she would edit articles that AI made. She said there would be so many mistakes! Whats the point of having an AI write if it doesnt even make your job easier 😂 Great video!!

[@nthelight079]

I now feel like someone needs to create a series of children’s books with your dog. Those images were so cute!

[@amiamarylis]

I am an exercise scientist and my friend is a ballet teacher and examiner and also trains other teachers. She introduced me to chat GTP, and we played around with it asking questions from our areas of expertise and the answers we got were incredible. I could not believe how good it was.

[@CanisLupusBruu]

Very nice coverage of this topic, really great work. Thanks Ann!