Should Artificial Intelligence SCARE You?

[@OwenJonesTalks]

Please like, subscribe - and keep this channel going here: www.patreon.com/owenjones84

[@thealternativeview2692]

Sooner or later, Owen, we are finished as a species. AI taking jobs will be the last of our worries. As soon as AI develops superintelligence, and the minute its goals do not align with ours we are in big trouble. Read the international bestseller Life 2.0. Brilliant and terrifying.

[@proe1]

I think this subject is worth a deeper discussion for Owen to tackle, as I think the affects of AI and how it is about to displace a lot of jobs is going to be profound, and it's going to happen closer to 10 months, not 10 years. I'm seeing it already in my job as a technical artist. The profit motives driving AI, the ridiculous pace of the technology, and the massive and currently under-discussed benefits that AI offers the technology companies that are creating it and pretty much every industry that can harness it is going to create massive problems for society.

[@dianacasey6002]

The one question I can’t get an answer to, is if ppl have no jobs and therefore no money, who buys the stuff that is produced?

[@ChrisJ294]

Man says it's just predictive text, but then says it can put opposing facts and summarise and we don't know how it does it. Make your mind up.

[@thenarrativecraft]

Fiction editor here! Proofreaders especially are worried, but all the heavier kinds of editing (developmental editing and stylistic editing) are somewhat subjective, and many people want to build relationships with human editors at the end of the day. I think our industry may shrink, and AI will complete some tasks, but that need for human connection and perspectives will remain.

[@brennuvargr4638]

Aye. I'm a proofreader and it definitely is worrying. I've only really just got my foot in the door, too, so it'd be a shame if AI took over soon!

[@thenarrativecraft]

@@brennuvargr4638 Even light copyediting isn't going anywhere soon, and my experience is that editors and proofreaders are very friendly, generous with their time, and happy to help one another out, so welcome aboard!

[@brennuvargr4638]

@@thenarrativecraft That's great to hear! Thank you, dude. :)

[@RaxLakhani]

09:11 Owen asks Michael about using ChatGPT to write fiction. Michael responds sating that the current text limitations of ChatGPT prevents long novel formats from being generated. Actually, this isn't true. You can keep prompting the programme to "continue" and the text will carry on being written. There are a number of people who are using ChatGPT to write books based on detailed prompts being entered in. The quality may not be award winning but the results are very impressive. Ask CHATGPT to write an episode of EastEnders or an episode of Friends and you'll get something back which is fairly detailed. Obviously, the more directions provided will result in a better script. Ask ChatGPT to write a poem about any perwon or set of circumstances and the generated response is absolutely brilliant. You can ask it to be written in a certain tone or using the style of a particular author and it does that too

[@riprap007]

Unfortunately Wooldridge is miss informing here on many things, such as LLMs size (there are many different models) to ChatGPT is 'just a sophisticated' auto correct. This is far too simplistic, as was all of his commentary. The conversation about AI is vital at this particular moment, partly because of the exponential development we are currently seeing, but also due to many of the tech CEOs ambition to create AGIs, which are a different magnitude to the narrow AIs we are currently seeing. One further, very annoying aspect of Woolridge's contribution was to not mention the the way AI has already badly affected society, one particular way is the algorithms inherent in recommendations on social media, and their creation of bubbles. I wanted to appreciate this, as I blog about AI, but I found it mostly annoying and unhelpful.

[@tuvaaq]

Absolutely but to be fair to the Prof. I don't think that was a bad analogy given this is a largely political channel and some if not many viewers may have no understanding of AI. 15 mins is not a lot of time to delve into the topic any meaningful way without losing the interest of the audience.

[@riprap007]

@@tuvaaq Fair enough

[@antonymossop3135]

I think the analogy is reasonable, the underlying approach is still a (notably inefficient) maximum likelihood response. I realise that source material quality can be reweighted, but it's still a glorified _vox populi_.

[@kevinwells768]

Well said. Over simplistic and not seeing the wood for the singe gpt tree.

[@Spiggle.]

The legend himself John Carmack is fairly positive AGI is not all that far away (some years) and to add to that, isn't openAI showing off a robot with GPT4 this summer? Sorry to this guy being a professor, but I'd trust John Carmack at least a bit more than him.

[@benfisher1376]

We are already in a world where people interact less with each other face to face, and date less , and are less connected to whats happening on the street because they are glued to screens. We are on screens half the time. AI is worrying because it would definitely be used by people at the top for their benefit. It would also contribute to society becoming less connected, more lazy etc

[@ecos889]

Probably not. but what should scare us is who owns it and what they use it for.

[@czero1243]

I feel your guest understates the affect on jobs these AI will have. Although machine learning is a better description of these tools it is already impacting visual design and creative processes. If it develops and can write a book in the style of a famous author how will you tell it isn't written by a feeling human. And he mentions efficiency but if large corporations can produce adverts, documents and many other things without the cost of humans do you really think it won't affect jobs. There will be just jobs for a limited few.

[@hanvyj2]

He's hugely underestimating the impact of these systems. It can't tie a shoe lace? How many jobs does that represent? These kinds of AI could replace managers, software developers, many doctors, writers, architects, engineers, project managers, customer services, financial services. It could manage accounts, send emails, answer phones... Basically office jobs. Which is a big part of our economy.

[@gianni_schicchi]

Ai is a better programmer than most of my peers already. I can give it a small sample of code and a sentence or two how I want to expand it and it can basically finish the app. It's like handing it two circles and asking it to draw an owl and it does in any style and language you'd like.

[@skyblazeeterno]

The concern isn't AI taking jobs the concern is will the people displaced be supported will they be able to survive. We have a system based on employment and we need to think of a future based on unemployment

[@nickjanczak9665]

Not necessarily AI, just the people who shove it in our faces and tell us it's the future and it's progress, it's those people and their motives as well as their lack of self-awareness that scare me.

[@italsounds001]

AI is able to reference multiple sources and put weight behind certain positions, for example the populist position is likely to get more weight than an opposing view, due to the amount of data that supports it, but it doesn’t necessarily mean that is the best answer, and it can only respond using the data available to it. The datasets being used are not all of the internet, or even a fraction of it, the owners of data platforms and the AI system owners decide what is available, it is easily manipulated to suit a certain outcome. If you look at how AI is being tested now, it has all sorts of useful and equally potentially troubling uses, especially where logic is the primary requirement, for example it can create a program to do a certain task, based on a description of what you want the program to do, analysing program languages, rules and script, it can come up with something usable already. I also wonder if people realise how much it is already impacting our day to day, the trick with AI is that you don’t know when it is being used, so you think it isn’t!

[@cdean2789]

Just magnifying the s-#@ out there

[@garyboyle695]

Driverless car! How many times recently has your phone, laptop, work computer etc frozen or needed rebooting, What a scary thought.

[@gatewayservices]

Not very informative Owen, didn't explain how many jobs have already been lost to AI, with some sectors like marketing being fundamentally altered. Also, centralised AI representing a further shift towards global subscription models for accessing knowledge and resources, rather than ownership, nor did you mention that AI constantly ignores copyright law, and that ChatGPT is only one of these such models - where graphics and audio creation (and soon video) from text prompts are replacing low-medium level workers involved in these tasks. But most significantly, is you didn't talk about the cost of access ... and that is the scariest part of all. When all these AI are priced in dollars (or perhaps worse, when they end up being created within authoritarian countries under government control), then who can afford access, and who is allowed access, becomes a substantial concern. The professor's point that these tools are blending themselves into ordinary computer, such that we begin to forget they are AI, is very important. Imagine every day tools becoming restricted in this way - that is the impending reality with any centralising processing system.

[@matthewtrow5698]

🤔 - "it's simply coming back with what it thinks is the most plausible response..." So, what do we as humans do, when asked a question? ... just putting that out there ... 😆 Do we consider responses that are implausible? Are human brains not trained on "large language models" - in terms of learning speech? Surely we are? The books we read, the movies we watch, the interactions with other people - all of these are a model. And yes, clearly those at the forefront of these large language models are intricately aware of this fact. The difference is we can also think in terms completely abstracted from language. It's so hard for us to really measure that - the other inputs aside from language. The myriad of inputs. That humans are capable of determining facts based on what they smell, as one tiny example. Pheromones etc. That facial expressions and body language are probably more potent and important than the spoken word. "It can get things wrong - a lot" 🤔 ... and humans don't? "The next guess when given a prompt" 🤔 ... and what do human minds do?

[@matthewtrow5698]

I'm rambling here, but if I had to come up with one modern analogy of where Machine Learning is at now, it would be to compare it to a Virtual Reality experience. We can shove some heavy goggles on our face and move around in a virtual world, but ... There's no depth of field There's no smell There's no touch There's no force - no physics. That's a useful analogy for me, as to where we are currently at with "AI" - it's still a parlour trick - a very very clever and convincing one, but a trick none-the-less.

[@nitsujism]

​@@matthewtrow5698 I suspect that the elephant in the room is that a lot of what human beings do is a parlour trick too.

[@a-rod6336]

@@matthewtrow5698 You're correct. I click off this video almost as soon as he said it's not thinking because there is no scientific way to test for sentience. It's not something you can make scientific claims about. Anytime someone says something isn't really thinking they're just doing pseudoscience.

[@davidsanders5652]

To be honest I would welcome ANY intelligence in charge of government. I've seen none since Wilson.

[@kevinfletcher1999]

A.I. could be far superior in governing a country and precisely why politicians would not allow it.

[@VentureHolly]

8:38 deadass sums up the UK education system from 5-16

[@hanvyj2]

"just a glorified auto complete" is doing it a disservice I think. Yes, that's what it is, fundamentally. But a lot of emergent behaviours are much more complex and have much more impact.

[@FSMDog]

Like with any technological innovation, the effect depends on how it's used

[@Urdsong]

I think it was Einstein that said if he had known how his work was going to be used he would have been a locksmith instead.

[@robertpetre9378]

As an artist myself, this is something I often wonder about will AI. Replace me when I just remember that my art comes from life experience Something A.I. can never really understand.

[@mawkernewek]

Can you write something about dahlias in the style of Owen Jones? Certainly! Here's a short piece about dahlias in the style of Owen Jones: "Dahlias, those magnificent blooms, bursting forth in a riot of color and complexity. For the uninitiated, they may seem frivolous or merely decorative, but in truth, they are symbols of the resilience and beauty of nature. Their petals, like miniature kaleidoscopes, reflect the full spectrum of light and life. They stand tall and proud, each one a unique creation, a testament to the power and creativity of the natural world. Yet, like all things in our society, dahlias are subject to the whims of fashion and commerce. Their worth is measured not in their inherent beauty or significance, but in their market value and popularity. We must resist this commodification of nature and celebrate the dahlias for what they truly are - a wondrous manifestation of the earth's creative power, to be cherished and appreciated on their own terms." Can you write something about dahlias in the style of Liz Truss? Sure, here's a short piece about dahlias in the style of Liz Truss: "Dahlias are a true British treasure, representing the rich history and culture of our country. They have been cultivated in our gardens and greenhouses for centuries, and continue to be a popular choice among horticulturists and enthusiasts alike. But dahlias are more than just a symbol of our heritage - they are a valuable asset to our economy. As we continue to promote British exports and strengthen our trading relationships around the world, dahlias represent a prime opportunity for growth and prosperity. With their vibrant colors and intricate shapes, dahlias have the potential to capture the imaginations of consumers across the globe. By investing in research and development, and supporting our local growers, we can ensure that this iconic British flower continues to thrive and flourish for years to come. So let us celebrate the dahlias not only for their beauty, but for the economic opportunities they represent. Let us embrace our heritage and look towards a bright and prosperous future for our country."

[@lucasprobably]

A counterpoint I have is that I cannot cook a meal, ride a bike or tie my shoelaces because of my disabilities, but of course I am still human, so why would an AI system have to do those things to be human?

[@leesteggles6367]

Already scary.. and it has already taken job's..

[@mawkernewek]

Can you write something about cats in the style of Owen Jones? Sure, here's a short piece about cats in the style of Owen Jones, the left-wing political commentator and author: "Look into the eyes of a cat and what do you see? A creature with a sense of self-possession, an independence that we humans could only dream of. In a world where power is often abused and hierarchies oppress, cats remind us of the importance of autonomy and self-determination. But let us not be fooled by their aloofness. Cats are also capable of great affection and empathy, as any cat owner will tell you. They are a reminder that even in a society that prizes individualism, we must also value our connections to one another and show compassion and solidarity. In a world that is often cruel and unforgiving, cats offer us a glimpse of the kind of society we could build, one that values autonomy and compassion in equal measure. Let us learn from these majestic creatures and build a world that truly reflects their spirit."

[@russellbrown8760]

Top drawer Owen! Thanks!

[@Marmite695]

Not at all! AI is intrinsically flawed, since it is a human construct. Thus, having been designed/defined/developed/derived etc. by humans, the flaw in AI is integral.

[@nitsujism]

Not when it becomes smart enough to create superior versions of itself.

[@Marmite695]

@@nitsujism Yet the original - with the intrinsic flaw (having been designed by humans) will perpetuate that flaw. I wonder if, should this AI being identify this flaw, it could eliminate the flaw without destroying itself?

[@hanvyj2]

Why is this the case?

[@hanvyj2]

Also, not sure why it really matters. If we are "flawed", and we create an AI, and for some reason like you describe it still has the same "flaw", it can replace a human worker. The economic and societal impact would still be massive.

[@vulcwen]

Before watching: Yes it should scare you, but just a little bit, just enough to know you need to be aware of what's going on and be ready to pivot when the situation arises. If you're too scared it's going to control you and that'd not be something we'd want either.

[@spacechannelfiver]

Betteridges law of headlines. The answer is no.

[@craigsanderson4330]

I cant for the life of me understand the path we are on. Technology for technologies sake? What world do we live in when we have no purpose.

[@skyblazeeterno]

Do you think work is your only purpose? If so that is sad

[@hilaryporter7841]

What does AI think about political ideology and about human morality? Given the blood thirsty history of mankind, sprinkled with bursts of intelligent enlightenment, if we we were using AI thinking it could guide us into the future, surely it would take us back to barbarism, because that's what it has studied most.

[@riprap007]

It depends on the LLM being used, ChatGPT4 for instance has been artificially barred from political discourse, you'd have to use an open source model, such as Open Assistant for political discourse.

[@kevinwells768]

It doesn't have a position on politics, but will regurgitate things seen on the Web.

[@nickfletcher4132]

Will we end up wiv Arnold Schwarzenegger? Will he be back?

[@michaeladkins6]

It will be Govenor of California?

[@cdean2789]

This could put politicians out of a job.

[@jonathaneffemey944]

Thanks for posting.

[@katherinecollins4685]

Interesting video

[@delphineblackburn3944]

How come it has been used in Switzerland to replace care assistants.?

[@VentureHolly]

Imo we’ll see a lot “AI inspired” art

[@Rorykdy]

Jaron Lanier would defo be a great guest for A.I episodes.

[@delphineblackburn3944]

He is selling his stuff. They are not bragging!

[@moshemankoff7488]

Thanks!

[@matthewtrow5698]

Not yet ...

[@matthewtrow5698]

We can ask ChatGPT: "As an AI language model, I don't have feelings, so I cannot experience fear. However, I can provide you with some insights that might help you answer this question. Artificial intelligence has the potential to revolutionize the way we live and work, making many tasks easier and more efficient. However, some people are concerned about the potential risks of AI, such as the possibility that it could become too powerful or even outsmart humans. While it's true that some AI systems have the potential to cause harm, it's important to remember that AI is still in its early stages of development. As researchers continue to work on AI systems, they are also working on ways to mitigate these risks and ensure that AI is developed in a safe and ethical way. Ultimately, whether or not AI scares you is a personal choice. Some people may feel more comfortable with the idea of AI than others, but it's important to stay informed about the latest developments in the field and to think critically about the potential risks and benefits of AI."

[@eternally66]

Everyone should be wary. Where is the legislation to control AI and the ones creating AI GODS!? Isnt using AI to write for us called plagiarism? Who wants to stop creative human input!?

[@mikeharvey9811]

Too scared to watch this….big brother on steroids. Barb

[@cdean2789]

When I was a kid in the 60s, we were taught that due to computers we'd have a major problem; what to do with all that leisure time. 🤣

[@gianni_schicchi]

This has already aged poorly. The answer is yes if you like having a job

[@robertmaitland09]

An AI government would be interesting compared to a no intelligence govt.

[@robm5358]

Should you be scared of it? Depends what your job is. I Have worked in offices where much of the work was automated by software. So jobs like office work is in quite immediate danger. Driving jobs I think is someway off because of our legacy road systems. IT jobs are in quite present danger too, and finance, and countless others that are primarily based working on a computer. And finally Labour, physical work. Here we have the merging of robotic with with AI, stil quite a way off. Although New York city are currently trialing a Boston Dynamics developed Robot dog-cop. You should ve more scared of what will happen evenually when all the unemployed kick off.

[@kevinwells768]

Anybody job acting as a carbon based robot without creativity or decision making capabilities will go.

[@hanvyj2]

​​@@kevinwells768reative and decision making will go. It's targeting the creative work first, which has surprised people. Anything digital (most office jobs) especially creative ones (problem solving like design, programming and software dev, engineering and writing) seem to be the first likely impacted. As confidence increases, it'll hit management and decision making roles.

[@kevinwells768]

I disagree. The easiest jobs to automate are those with routine procedures and a limited set of queries, hence stock control, finance and hr depts. They'll still exist, but you will just need a very small team of humans to handle the oddball issues that arise, so instead of 12-15 people you will just need 2 or 3 to handle these cases and keep an audit on what the ai is processing. We will still need coders and creatives the difference is that the machines allow a weeks worth of work to be completed in hrs. So, some overall job losses perhaps in the latter, but machines make mistakes, don't believe all the investments opportunity hype, humans still need to be the final arbiters , and take the wrap accordingly.

[@stephencalder1583]

This presupposes there is such a thing as objective truth about historical events. An AI book on the history of WWII for example would have different biases depending on whether it was from the point of view of the ruling ideas of eg Russia, Britain, USA.

[@nathanbradley4092]

As someone that is dyslexic this is brilliant for proof reading achemdmic work

[@treyquattro]

Real Stupidity is what scares me. Oh, hi conservatives! 🙂

[@WarrenPeaceOG]

What's your mum say, Owen?

[@Skylark_Jones]

I thought the most scary thing I saw a few years ago was a video of dancing robots.

[@philsophkenny]

🤔

[@amylou7991]

I wonder if sentient AI will be concerned about tying shoelaces or cooking an omelette?

[@jamietulacz7742]

Really interesting stuff, and slightly put my mind at ease. Not enough people are thinking about the societal consequences of AI imo, it would be interesting to have someone on your show to talk about that perhaps?

[@kevinwells768]

Lots of office jobs will go, HR and finance deps will be shrunk ten fold. LLMs can now be trained on a desktop with a gpu, stanford has already brought out a paper on this using chatgpt to train it.

[@Melissa.Garrett]

I have to say that there is no way in hell an AI could write a readable novel. Art of any kind has to be motivated by emotion in order to speak to the emotions in others, which this technology could never do.

[@kevinwells768]

That's coming. It won't win a Booker prize, think of chick lit.

[@ericangus2330]

I’m not so sure. As AI becomes more sophisticated it could perhaps access the randomness of human emotion and imagination - then maybe produce interesting art. It can’t at the moment but it is developing fast…..

[@czero1243]

This is just not true. It is already emulating emotional art because if you can analyse millions of previous emotional material you can write so similar to real emotion the viewer cannot tell. It is still early but it is already creating visual art that is getting close and how many years will it be before it can produce books, movies etc. It has the ability to accumulate so much it can convince our emotional response so sadly it will be capable.

[@hanvyj2]

You are simply wrong here. We already have AI generated art winning photography and art competitions. AI novels will be a thing in a few years. There might be a place still for humans with very specifically creative novels that try something actually unique. But the vast majority of novels and art are not unique. It'll be the low hanging fruit first. Chick lit, romance, pulp fantasy - genre novels that follow rules/patterns more closely. But it'll likely eventually be most novels. When you can't tell the difference as a reader, it'll be a huge advantage to use AI. Bear in mind, most won't just be solely written by AI - if a human had a creative idea for something novel they could just ask the AI to write it. The bulk of the work would be done by the AI.

[@goingoutotheparty1]

Great stuff 👍👍

[@WarrenPeaceOG]

Computer coding seems to be something AI is very adept at as there are standard rules and protocols and procedures. I found the ChaosGPT story reassuring and funny: using AutoGPT - automated GPT, which you can give big goals to work on indefinitely - and telling it to sow the seeds of chaos. It was a damp squid ;-) It put out a couple of end is nigh tweets😹)

[@WarrenPeaceOG]

Having read all the comments: my biggest concern with AI is billionaires

[@MisguidedDive]

It doesn't scare me as much as human intelligence

[@steelic1]

AI? Not as much as AT(Artificial Trans) ...

[@Tvjunkieful12]

Great vlog, brilliantly presented. With just a few question you got professor Wooldridge to spill his wisdom and knowledge and explain in a very simple way something that most of us only know from sci-fy movies, and headlines and articles inspired by such. Having listened to this it was one question that popped into my head, I would have loved to have heard the professor's answer to, and bare with me cause I don't have the educational skills of professor Wooldridge: Would it be possible, by combining nano tech, mag lev technology, combining electricly operated chemical or hydrolic artificial muscles, ditto controlled and operated touch and flight, maybe even sound, sensors, to together with AI create a robot that could do hole-in-one everytime on a golf course, eternally get strikes on a bowling alley, and similar? I imagine the golf robot would both be able to calculate variable factors like wind, or lack of, what angle and force a ball would need to land in the hole, and that mag lev and artificial muscles would make it able to move like a human - only with greater precision.

[@rodd1000]

Maybe humans will actually begin to love one another when there’s a new bigger common threat. Could be the break we need. 🤷‍♂️

[@skyblazeeterno]

I think the threat of ai is that it will free people from work but far far too many people have their sole identity linked with work so it's scary

[@annlove2098]

M>

[@annlove2098]

M