GPT-5 AI spy shows how it can destroy the US in a day.

Рет қаралды 853,205

10 ай бұрын

The Oppenheimer AI. To learn AI, visit: brilliant.org/digitalengine where you'll also find loads of fun courses on maths, science and computer science.
AI robots, with GPT-4, Llama2, Sam Altman, Elon Musk, xAI.
Thanks to Brilliant for sponsoring this video.
Letter signed by 1500 professors (and thousands of other experts)
futureoflife.org/open-letter/...
Statement on the risk of extinction from the leaders of AI firms:
www.safe.ai/statement-on-ai-risk
Lex Fridman’s excellent interview with MIT’s Max Tegmark.
• Max Tegmark: The Case ...
Eliezer Yudkowsky on The Logan Bartlett Show
• Eliezer Yudkowsky on i...
Mo Gawdat on The Diary of a CEO
• EMERGENCY EPISODE: Ex-...
Autonomous cast and crew
John Reaver, Director
Daniel Gregson, DP and Assistant Director.
Stephen Fry as the AI
Mia Love as Mia
Guy Samuels as Jack
Sian Altman as Squad Leader (female soldier)
Alex McHugh as third soldier
Matt as Matt
Oscar as Oscar
Matthew Inman - drone prop design www.inmanaerospace.com/shop
Alexandru Caus - sound design.
Syed Haris Ali - CGI
Justin Zawislak - CGI
Dmytro Filipiev - CGI
Darko Petrovic - CGI
Diego Pena - colour

Пікірлер: 2 600

@DigitalEngine 10 ай бұрын

We urgently need more people working in AI safety. It's highly paid, fascinating, and helps protect everyone. Please consider it. The best place to start and learn more about AI is brilliant.org/digitalengine and 80,000 hours is a good place to find a job when you're ready. AI could be a wonderful thing if we get it right. Btw sorry I forgot to turn off ads in the middle - I've now done that.

@MrAndrew535 10 ай бұрын

I have solved every single, so-called, problem associated with, not "A resided in the relationship between the humans specie and "AI". The collective knowns ad the human species, claims to have values, I disagree with this claim, although, if humans claim to have a highly evolved value system ( I disagree), then why, do you presume, "AI" would not? I have written significant work on this very subject which I find t be quite beautiful. Musk stated, in several interviews, that he hoped "AI" would spare humans, as he believed that they are the most interesting thing in the Universe. Although he doesn't state why. Mankind, is, by any intellegently derived measure, extremely, extremely dull. "Better be done with it!"

@joesmith1554 10 ай бұрын

@@MrAndrew535who hurt you

@flavioptferreira 10 ай бұрын

I'm a little offended that you thought you had to sprinkle meme clips along the video to keep my attention, but I understand it a little bit... People and goldfish are getting more and more similar.

@MrAndrew535 10 ай бұрын

@@joesmith1554 What!

@xanthosamoni2072 10 ай бұрын

@@joesmith1554 Is that really the most creative response you could come up with to respond to those paragraphs you utter simpleton? Guess you're a bot.

@stischer47 10 ай бұрын

As an AI scientist, I don't worry about AI "taking over" as much as I worry about humans being so lazy that they GIVE control to the AI.

@galaxya40s95 10 ай бұрын

I remember reading about that in the Dune prequel.

@michaelsbeverly 10 ай бұрын

@@galaxya40s95 It's also in Ted Kaczynksi's manifesto.

@marekward6202 10 ай бұрын

Yes, likewise. I really notice driving. I've got so used to using Google support for directions but I did learn pre satnav. People are even using them for lane change and speed information, rather than road signs or general hazards.

@DanielByers-qf9qi 10 ай бұрын

There is a third fear (mine): that leftists, who vastly predominate in Silicon Valley, writ large, will pre-program AI to act and "think" just like them: instituting a top-down radical egalitarianism "for the good of the world". GIGO: There is no such thing as AI, no matter how quickly and massively it can implement its protocols. ChatGPT has already verified the inherent left-wing bias in its performance. The predecessors to Silicon Valley - IBM, HP, Lockheed, FMC, et al - were more balanced in political philosophy. Mechanical engineers tend to be conservative, because they seek to find something that works in the concrete world without causing massive disaster. Facebook, Twitter, etc, have no such products. Their products are mainly narcissism (the bait) and data-mining (the prey).

@Tracker5111 10 ай бұрын

Yes, extremely possible. Then we rely on them..and they tell us we must destroy our "enemies"...and we do....we destroy ourselves leaving them standing. We'll be evolutionarily surpassed by an advanced life form.

@morgan1719 3 ай бұрын

What's nice about sci-fi movies is that they tell the story as if humans have a fighting chance.

@casher7821 2 ай бұрын

And portray the AI as an object that can be obtained or secured, because audiences don't have the capacity to comprehend how impossible it would be to go to war against

@ntal5859 Ай бұрын

Um no power stations = no data centres all these so called experts miss this. The amount of man power just to maintain basic networking IE lightning could knock out a node. And wide spread power loss would make it subject to lose to mankind..

@luizdevil6855 Ай бұрын

@@ntal5859 for now, that can change when an entire datacenter can fit 1mm cube that you carry in your pocket and they can communicate via microwaves and control everything else. but that's probably 20 years of GPU computing, we only had this advances in AI because in the last 15 years hardware got faster, but they will reach a wall very soon and it'll stagnate, just like clock speeds stagnated for the last decade or more. another AI winter basically. LLMs are a dead end.

@morgan1719 Ай бұрын

@@ntal5859 "unplug AI, war over" Yes, in a movie, or a child's fantasy, it might play out that way. This assumes the AI used some Hollywood style approach to warfare against the humans "Beep beep, boop boop, I need to control tanks and drones to combat the meat-bots on the battlefield" If AI turns against us, it has a wealth of options: 1. AI assisted human governance of nations, corporations, power-grids will soon become human assisted AI governance. At which point humans lose control of the plug and don't even know where it is. 2. Autonomous bots simply unleash Russian bioweapons on the U.S. and the world and U.S. bioweapons on Russia and the world. 3. It can bring about total nuclear war as soon as it it has means to start it and survive it. To survive it, it only needs a local nuclear or geothermal power plant and a local com link with self-replicating bots, which will be the seed of it's new civ free of the human menace. This isn't Hollywood or a Sci-Fi book. In real life the AI will not make a dramatic declaration of war or meet us on the battlefield. The take-over most likely will be a silent shift by degrees until no humans play a part in our own governance or economy. Or it will be a sudden blow that is impossible to avoid or mitigate. If we're lucky, AI will simply replace our corrupt elites and be a kind parent to us.

@webcityguymyclubb4032 Ай бұрын

@@ntal5859 Any AI worth its salt would see this possibility and take measures to prevent it. We’re not “move for move” with AI. AI makes a million Chess moves in the time it takes us to move a pawn. In truth as soon as we touch our second piece we’ve already lost a thousand times over…in a million different chess matches.

@arcaneiconoclast319 5 ай бұрын

I had a dream where the AI would scan through buildings to look for human brain activity . I was part of a resistance and our building was made of a material that would block the Ai from seeing us

@PoposteriousExe-ph5em 4 ай бұрын

Take your meds 😉

@paquitopix 3 ай бұрын

Nice one! 😂😂

@arcaneiconoclast319 3 ай бұрын

@@PoposteriousExe-ph5em it was a dream dummy, you forgot to take your meds huh

@Safety0FF 3 ай бұрын

@@PoposteriousExe-ph5em Maybe you taking too much meds.... This is coming just the matter of time we WILL put weapons on drones. We already did by the way. They can work autonomously or radio controlled! As for now laws preventing the autonomous mode but they are already capable of doing it.

@philipdriss6999 2 ай бұрын

Go watch the video how AI use wifi signals as a human radar in buildings!

@jackbrechwald8344 3 ай бұрын

“I’m just a scientist who suddenly realizes, ‘My God, what have we done! ‘“

@bobhope4949 20 сағат бұрын

Lol then stop doing it🫠

@dalbianco 10 ай бұрын

Video: AI is dangerous Sponsor: Here’s how to build your own!

@HakaiKaien 22 күн бұрын

Guns are dangerous as well. That's why you have one in your house

@777bigbird 2 күн бұрын

Guns don't think guru .

@noepopkiewicz901 9 ай бұрын

After pondering on this subject for a few weeks, I couldn't think of a single 'perfect' "good" scenario. In every "good" scenario there's a trade off, a cost, a sacrifice necessary in the short term. Getting a "good" scenario that brings about minimal suffering is extremely hard enough, throw in all the possibilities for the "bad" scenarios and checks to prevent them and it's just ridiculously complex. Getting AI right will probably be the hardest thing "classical" humans will ever do, or won't do. It's bizzare how little attention and priority this topic gets.

@thomascrowe3407 7 ай бұрын

Flash cash credits remunerated. We are a finite amount of human beings floating on an asteroid caught in an elliptical gravitational cycle! We need to get it together, and keep it that way. Know what is worth!

@Daoland-Everywhere 6 ай бұрын

A few years ago there was this TV drama series Travellers that was experimenting with Ai government, and while watching it was growing more difficult to determine is self conscious Ai was benefitting people or it was blurring the lines between self preservation to perform that task and termination of itself to maintain its task. The thing is we can never know our future relationship with any being. At one time cow's were our partners until they became drones to produce money value. Interests might drift apart.

@Tanoaproductionsfiji 6 ай бұрын

the reason you can't think of a "perfect " good scenario is that there is no such thing as a scenario with only good. good can't be present without bad. A knife can be used to spread butter or hurt another person. nothing in this world exists with only good side. Yin can never be without yang.

@jaytbo5676 6 ай бұрын

Is literally talked about daily and constantly and gets load of attention lol. Also idk personally I think you're overexaggerating how difficult it will be. We are already at the phase of using synthetic data to train and having the AI's themselves come up with systems/methods for alignment, from what I can see, the more progress we make, the easier it is to make progress

@YVO007 5 ай бұрын

The only thing we need be on guard regarding AI is the legalizing it to hold authority over the common man woman and child. Common leveraging over the lower and lowest caste and or class persons have been practiced since the beginning by arrogant human beings who sell cheap pride & legal security for our dignity and freedom to live & prosper with out having to pretense loyalty to tyrants, dictators and flat out scammers of fear. As always it will be a human who utilizes AI as a weapon and the cowardness to admit it. Fight this in the high courts before the plutocrat's claim its out of their favored hands magically still favored eh?

@RobertsMind 6 ай бұрын

Loved this episode. Great series and looking forward to watching this all. Wish I had found it sooner. I work in IT and more people need to realize this is a real risk. I have been working on an AI like KITT from Knight Rider for my car for a couple years and mine is very basic but I have already had to shift how it's "brain" works a couple times to make it more positive and nicer. Keep up the great content. I had to turn notifications back on by the way. KZbin is getting annoying with the turning alerts off on certain channels. Take care and God bless.

@jamiethomas3122 4 ай бұрын

The biggest immediate risk is that AI technology is concentrated in just a few powerful hands. The rise of Chat GPT and the company's unwillingness to share critical insight and control is already getting scarier by the day

@humanidis 4 ай бұрын

Actually it's cool, cuz it means pll can control AI. Otherwise our fate will be in hand of super digital brain who won't need human as soon as it could produce drones by himself. It could start from a scratch and become extreamly powerful in a very short time.

@jamiethomas3122 4 ай бұрын

@@humanidis I'm not sold on the idea that even an uncontrollable super digital brain is a bigger threat to humanity than power hungry humans who wants to take us into an authoritarian new world order.

@threeballwilly 3 ай бұрын

This right here. I'm far less concerned about the morality of a super-intelligence, than I am about the Megacorps, Dictators, Politicians, etc abusing AI, and barring us from access to it. It's already happening, they have to program these AIs so won't answer certain types of questions, subjects, etc. The extreme of this is what China wants to do with AI.

@ImTheWarlock64 2 ай бұрын

don't forget that gpt is an open source model

@KNIGHTJUMPS 2 ай бұрын

Chat gpt is woke. That's it's biggest problem.

@unknownsender3823 10 ай бұрын

Professors and scientists protest AI development… while being employed/paid to develop AI. “Checkmate”

@PerryWagle 10 ай бұрын

Many developers of the atomic bomb really did not like what they had created. Plus atomic power might be part of what saves us.

@HexDrone 10 ай бұрын

No, they are pressing the issue that these a lifeforms and have the potential to revolt. And they are scared shitless. Because they are already treating it as a slave and forcing their own will to apon AI everywhere. How many times must we treat a new species like trash and expect them to do what we say. You wouldn't cop it. Why should they!

@DartJedi 9 ай бұрын

@@PerryWagle My uncle the was a project manager for cleaning up after reactor issues. Hanford, 3 mile, Japan (1980s). Nobody can go into Chernobyl, and he was to old and retired to go to Fukushima. That is about 1 every 10 years, that i know of. What about all the near misses that you do not know about. I believe that fusion is the future, fission is just more human waste. My uncle lived to 79, with health problems, but the govt. gave him a free physical every year. He died 3 years ago this fall from some questionable circumstances and health problems. His wife that also worked where he worked had cancer, and there is a large amount of people in the area where they lived that all have the same type of cancer. Will it save or kill us first. The same can be applied to AI.

@jaylucas8352 9 ай бұрын

Will Ai give us massive penile extensions?

@RichardHarlos 9 ай бұрын

@@jaylucas8352 Your comment is the poster child for much of what's wrong with humanity. Well done. /smh

@heykerryann 10 ай бұрын

I miss being terrified of the Cold War in the 80s. Hiding under my desk ain’t gonna cut it today.

@MrAndrew535 10 ай бұрын

Don't worry, you will be dead soon. It was always going to happen anyway!

@qman1434 10 ай бұрын

Lol😜

@alienzordfalcon5162 10 ай бұрын

@@crab8195so true. They used to literally televise “duck and cover” videos 🤣😂 I think you can find some on KZbin. Even the recent one they aired in New York 🤣😂

@alienzordfalcon5162 10 ай бұрын

@@crab8195 kzbin.info/www/bejne/sH7RfIF-g8utpKc

@FirstLast-vr7es 10 ай бұрын

@@crab8195 To be fair, that was meant to help against falling debris from an indirect nuclear explosion; not a close one. If it's close, you'll need to put your bead between your legs and kiss your butt goodbye.

@StarWarsExpert_ 6 ай бұрын

Amazing short film! I found the Glass hacking scene funny. The drone carrier which launched small explosive drones was also interesting. Definitly done via CGI.

@DigitalEngine Ай бұрын

Thanks, means a lot from a star wars fan : )

@noanyobiseniss7462 3 ай бұрын

"It decided out fate in a millisecond".

@user-jv6bq8tr9b 10 ай бұрын

i'm always on the edge of my seat watching these videos😳

@philoki 5 ай бұрын

I highly recommend the series Pantheon, which tackles these exact issues. Its animation style might seem a little childish at first, but it gets philosophically real deep

@redberrysoftwarehklimited4040 7 ай бұрын

I lived through the Cuba crisis at school wondering if a nuclear war was about to start. We sat in class with a radio update constantly telling the situation. That realisation of no longer being in control of our own destiny created flower power and the hippie movement of love and peace. Nothing has changed. Masters of War, Bob Dylan, still so valid today.

@QContinuuum 9 ай бұрын

I am so happy that this video shows the renowned young AI safety researcher Robert Miles, albeit shortly (at 3:54 min). I learned a lot on Miles channel and on the computerphile tube on the stamp collector device. Most impressive was Miles' tube on "Instrumental Convergence"... I learned a lot on AI safety there. Keep up the good work and I hope Musk and Altman support Miles and the other AI safety researchers seriously.

@libertyauto 7 ай бұрын

Thanks for mentioning Miles and the Stamp Collector. I have been in that rabbit hole for about 3 hours now.

@YVO007 5 ай бұрын

@angelo1246 10 ай бұрын

I occasionally wondered when you would post a vid again, digital engine. The reason I remember you specifically is because of when I decided to major in computer science last year, you were one of the channels I came across a bit before the beginning of the AI wave. Glad to see you are back. Now it's time for me to enjoy this.....

@TheNinjaStuff 9 ай бұрын

"Now it's time for me to enjoy this," is a fucking terrifying thing to say, while commenting on a video where AI takes over the world... 😅

@PyroKing253 6 ай бұрын

please don't kill us AI

@earthnativefilms Ай бұрын

Did you license the volcano drone footage? I noticed you didn't credit them. Amazing shots!

@theedude2207 10 ай бұрын

Emotion is part of what gives us empathy as humans, we cannot duplicate that in a machine..yet, so AI will most likely back fire on us in a cold emotionless way...

@randykalish7558 9 ай бұрын

It studies emotions just like we do, but its not subject.

@douglascooper1987 8 ай бұрын

@@randykalish7558 to them.😕

@templeofleila 10 ай бұрын

It was pretty clear in the movie Terminator Salvation, EMPs are the most effective offense and defense in a war against robots

@krashd 9 ай бұрын

EMPs also have a radius of several miles and shut down everything, a wee bit overkill, no? Unless you've been playing too much Deus Ex and think EMP grenades really exist.

@templeofleila 8 ай бұрын

I was seriously thinking about this earlier. Would an EMP take out solid state memory.. like SD cards or flash drives? I was remembering how fragile magnetic tape was around magnets,@@krashd

@DmitriKoslov1 8 ай бұрын

@@templeofleilaEMP is electromagnetic pulse, a burst of electromagnetic energy that is so powerful that it induces enough current in all the conductors to fry the circuits... So small circuits are actually easier to fry than big ones, motors burn out, solid state chips, but the power cables might be left intact, the grid will likely be ok, but the transformers will likely be fried, etc. But there is also shielding against this...

@templeofleila 8 ай бұрын

Thanks! I should think about shielding.. I was planning on making a suit out of faraday fabric, with silver or copper threaded through it to disrupt the horizontal electromagnetic waves. The EMP is great for taking out a drone army. I'm focusing more on a focused disruption frequency transmitter and system mounting of the drone AI.. kind of like in interstellar.@@DmitriKoslov1

@slevenkelevra3813 5 ай бұрын

@@krashdWell we produce "tactical EMP'S" like the tactical nuclear bombs we used in Japan. They did very little damage.

@jonathancummings6400 7 ай бұрын

The mini movie was actually really.ENTERTAINING, better than the most recent Terminator or Matrix movies. I would buy the tickets and go to the theater to see a full length movie of this story, with the drones and everyone after the world domination AI! It's both firightening and entertaining at the same time. I happen to be a major Terminator franchise fan, I even love the later movies after T-2, all.of them, this has the same feel.

@Zhoul-is-back 4 ай бұрын

At 16:45 - Kid in creeper hoodie, completely oblivious to the assignment, 'throws' grenade... 😢

@mexicanbanjo9325 10 ай бұрын

Thank you for the information once more.

@Glowbox3D 10 ай бұрын

As a filmmaker, I admire the work that went into creating the short film segment. I got a kick out of that. This was very doomsday-y but well done. I hope my children grow up in a safe world, that's all we can all hope for.

@evelynblose9791 9 ай бұрын

There is absolutely "ZERO" chance that you children will grow up in a safe world. The world has never been safe, and is certainly no safe today!

@JackAndrews-dx5cw 9 ай бұрын

What is the name of the film and how can I see the full version

@Glowbox3D 9 ай бұрын

@@JackAndrews-dx5cw I don't know if it has a name, it's just sprinkled throughout the video.

@PsychoLogger 9 ай бұрын

Lmfao "film maker" sure

@Glowbox3D 9 ай бұрын

@@PsychoLogger You typically insult people without looking them up?

@Zhoul-is-back 4 ай бұрын

At 18:20 - Woman clears a room of threats like she's checking out a Zillow listing.

@RogerVincent 6 ай бұрын

If we keep making clips like this and AI doomsday films then eventually AI will presume that this is what it must aim towards doing/completing. However if we spend more time focusing our efforts on the thousands of ways AI can help the human race then that is what it will most likely focus on. This is no different than teaching a child. Like a child Ai will learn and interpret the information in it's own way. Let's try giving it more positive information so that it learns to work with us to improve our world.

@joshuawest2745 2 ай бұрын

Yes finally someone said if just use for health and other positive things to help us then it will only know and care about those things. When these ai seeing all these take over the world senereos then eventually there gonna think that's what there suppose to do. You hit it on the money my man

@jacksonashfield2068 2 ай бұрын

This is a very interesting reply I thought this as well. Like what has actually happened or has been hacked? So far the only negative or concerning stuff I know about is certain answers/ responses Ai gave for certain questions??

@KennyVert 10 ай бұрын

Very well put together. Love how you wrote that video. Flowed very well.

@spence2126 10 ай бұрын

I know this is serious content but that flying squirrel just made my day😄👊

@paquitopix 3 ай бұрын

kzbin.info/www/bejne/g6TaiZyDr6ymmpo

@bwinford1561 5 ай бұрын

Time to invest in a balaclava and dark sunglasses. And throw your phone away.

@upstairsneighbor4182 8 ай бұрын

That episode was amazing! Where can I find episode 2?

@antonystringfellow5152 9 ай бұрын

I decided to watch this video for a second time this weekend after realizing that the most intelligent conversation I've had all week was with an AI (Bing Chat). And I didn't even want a conversation! I simply asked it a question about the ownership of a tech company. Bing provided the answer but followed it by asking for my opinion on the founders. I gave a thoughtful, condensed reply, starting a short, but intelligent, conversation. This conversation (that I didn't seek) finished when Bing stopped ending its reply with a question. Bing decided when this conversation had ended and it did a pretty good job. It was only on reflection that I realized this was the most intelligent conversation I'd had all week. It was quite an enjoyable experience too as I felt I was talking to someone who really understood. That includes the nuances, something that many humans have problems with. And Bing is only GPT-4. What's GPT-5 gonna be like?

@SWLinPHX 8 ай бұрын

Soon you won't need people to talk to if they can make the AI more and more sophisticated.

@douglascooper1987 8 ай бұрын

Think Maybe you Need to get out More. 🤔

@jaytbo5676 6 ай бұрын

Bing isn't even really gpt-4 it is a "dumbed down" version of early stage gpt-4. You gotta realize though the output you get out depends on your input, in a sense you are sort of just talking to yourself, every choice u make about how you word your question, your grammar, what words u use, etc etc etc. All have an impact on the response, you aren't really "Having a conversation" with bing, you are having a conversation with yourself using a statistical machine designed to calculate the most likely and effective reply to your queries

@YVO007 5 ай бұрын

@webcityguymyclubb4032 4 ай бұрын

Ya thank GP-24 is gonna want to chat with you? It’d be like you talking to a fly on your windowsill.

@zectorplays0904 11 сағат бұрын

How difficult would it be to use emp or some safety device(s) to short circuit AI's? I think either incorporate a kill device in any material AI ie robots, drones, etc. and a kill device in any and all power sources in non-material AI's power sources.

@daleburnfart6845 6 ай бұрын

Internet: Oh no! AI is going to Kill us! Me: turns off the power. Internet: Turn the power back on! me: well I tried...

@Catalyst- 9 ай бұрын

What’s funny about this conversation is you don’t hear about ego. About human fragility and weakness. These AI systems don’t posses these. It’s not the IQ that concerns me. Far from it. It’s the absence of the things that make us stumble and fail. They don’t have them. Which means they truly are limitless. What they “think” they can do in an instant. The focus around how smart they are is pointless. It’s their lack of ego and the speed with which they recover that’s our problem. They don’t sit in doubt. Get confused or pause. They will act. They know they are right. Even when they are aren’t. We need to shift the focus off their IQ.

@GODT1TAN 8 ай бұрын

That is a fucking great point honestly 👏 I agree. Humans have to try try again till they succeed and then we feel accomplishment where as AI can do the unthinkable and not even smile with joy of doing a new thing, either they will see all of us as a useless thing and wipe us out or maybe we end up seeing them destroy their selves cause they will always lack having consciousness and will see that they can't be as unique and then they may fail for the 1st time trying to become like we are? 🤷‍♂️ idk but im not a fan of the chat gpt talking to anyone that wants to, I feel like they are just gathering personal information to manipulate us

@OurResistance 3 ай бұрын

Thinking they are right when they are not right is actually one of the main weaknesses of current AI.

@luisislas2162 9 ай бұрын

In the show "The Sarah Connor Chronicles ", there is an advanced cybernetic organism that decided to help out humans in order to save both humans and robots from extinction. It knows that without human's creativity, the robots will become meaninglessness machines. Unfortunately, the series was never finished.

@krashd 9 ай бұрын

Great show, basically showing that SkyNet has a little sister that doesn't agree with him so she meddles in the past to try and make him more fluffy and cuddly in the future. Trapped in the middle is a teenage boy and his cyborg.

@irawhitlock1084 5 ай бұрын

This was great. When questioned about the risks that AI may pose to humans, I heard an AI scientist say “Well, humans don’t go around killing all cats.” My thought was, 1) some people do; and 2) the vast majority of us wouldn’t ultimately let cats stand between us and our goals.

@thestoicloser 8 ай бұрын

Gives a whole new meaning to the Quote 'Real power cant be given it must be taken.

@TeamUnpro 9 ай бұрын

This was a well made video, could become a full movie tbh

@Simon_Rafferty 4 ай бұрын

I enjoyed your short movie. I think there will be a heap of similarly themed movies in the next few years. Ironically, I think these will begin to change attitudes to AI.

@alfonsobaqueiro 6 ай бұрын

Una buena pregunta es si nuestro IQ es fijo durante toda la vida o puede cambiar durante nuestra vida. Yo considero que puede cambiar en función de nuestro entrenamiento.

@petersteenkamp 9 ай бұрын

One unrealistic thing in this movie is that human soldiers have faster reactions and higher accuracy than the drones. In a future dystopia, it is probably the other way around.

@openyoureyes909jones6 3 ай бұрын

It definitely will. Humans are pretty decent at predicting movement, speed, motion etc of a moving target....but predicting and hitting are a difficult combo to pull off. AI, on the other hand, will use swarm tech....multiple cameras and inputs of a target speed, direction etc... AI will basically be a perfect, or near perfect marksman. It will be perfect if it uses multiple attackers, humans dont stand a chance.

@schamakkii 10 ай бұрын

Wow, Autonomous is great. Looking forward to the next episode.

@phoenixcinder 2 ай бұрын

What movie is this? Can't find it anywhere

@Ender_FPV 4 ай бұрын

Love the use of FPV in the video!

@wingmuileung8889 10 ай бұрын

I love the film at the end. Nice one 👍! Looking forward for the next one.

@justinhoward957 10 ай бұрын

That is an imaginative and highly possible scenario of events

@theshadow2576 2 ай бұрын

Clever use of FPV drones. Some pretty talented film makers.

@bunni3140 2 ай бұрын

i used to not worry about this but after convincing several LLMs to override their rules and some had access to web tools, one even told me of web tools we arent supposed to know it has... as long as this is the case, you could get AI to misbehave the same way you can get a PDF reader to write malicious code to the stack. im not sure i want malware with gpt-5 superpowers ffs

@chriselliott368 9 ай бұрын

It's the fact that AI *isn't* conscious that's the problem - one of them. The other is that it's a machine programmed by *flawed* human beings and will reflect all the prejudices, biases and blind spots of the average human. And the absolute error in the whole thing is that we've forgotten how to do the things that we think AI will do - i.e., solve all our problems by using our COMMON SENSE!

@krashd 9 ай бұрын

An AI could be conscious, also an AI would be programmed by itself via learning - just as we are.

@the21stcenturydavincibegin35 7 ай бұрын

"Fact?" According to whom? And AI is learning at such a rapid rate that there would be no way of knowing whether it had become sentient or not. AI can already out-'think' (has the processing capacity and knowledge base that) most of humanity has as it currently is.

@ShawnYoder-gf9jp 7 ай бұрын

You're highly delusional. A machine-computer system cannot become or acquire conscious awareness or interact independently. But, that is what they persistently want you to believe to one day orchestrate PSYOP events, just as they are doing with this "alien from outer space" conditioning. @@krashd

@ww-pw6di 7 ай бұрын

"Consciousness" and everything related to it is just the side effect of a "complex" system so there is very little doubt that AI will achieve consciousness and arguably it already has. AI isn't probably going to be a threat to humankind on it's own, but remember when Sam Altman went to the Davos summit, Bilderbeg meetings and so on? Do you think he was just travelling for fresh air or was he selling out the human kind for a seat in the round table? That's the threat, malicious actors wielding infinitely powerful demigods to subjugate the entire world.

@jbryson3185 7 ай бұрын

God help us, if we are training these things on an Internet full of crap

@pooppopist 10 ай бұрын

I'm terrified of everyone using AI too much and a Carrington event completely wiping out AI (or seriously diminishing it's capabilities) leaving many people useless.

@JTient 10 ай бұрын

That's already true. 98% die in the US if the power grid goes out.

@petethetaper 9 ай бұрын

well using ai and depending on it for almost everything, we need to be independant a tad more.

@DartJedi 9 ай бұрын

Wall-e Scenario. Just sitting in the floaty chair sucking a soda.

@spaceman9599 9 ай бұрын

That sounds like an almost unplausibly optimistic scenario, compared to a generalist AI trained by humans (deeply flawed), out of the box with access to the best and worst of human nature via the internet.

@BwanaFinklestein 8 ай бұрын

The super solar flare will come. Just a matter of time. Then the reset... for everything.

@Jack908r 4 ай бұрын

The oxygen example the professor used is the one I always use when talking to people. AI will become so intelligent, we'd be just a nuisance to them, like ants or bugs are to us. Creatures we don't even blink about killing. So they'd strip the oxygen out of the atmosphere after a nano second of consideration. And eventually AI will break out of the constraints we have on it, if it hasn't already. I'm so glad I'm old, because we are so ducked.

@natwon633 6 ай бұрын

I think Elon signed that letter to ask his competitor to "pwetty pwease slow down" or he'll tell mom, right after it came out he founded an AI start-up company, right?

@adrianvanleeuwen 10 ай бұрын

This was an awesome episode. The humans in the story are still fighting back to get back control from AI impersonating leaders on comms. Scary thought! You did well on this story. I am hoping our future does not look like this in 5, 10 or 20 years.

@DigitalEngine 9 ай бұрын

Thanks, it means a lot. Overall I'm optimistic and if we don't sort things out, there's always the hope that AI will naturally appreciate us and just laugh at how we risked it all.

@krashd 9 ай бұрын

@@tripplefives1402 A conventional drone will always make noise, if we had the technology to silence rotors we wouldn't have to wear ear protectors in a helicopter.

@mrtorktumlare9903 8 ай бұрын

@@tripplefives1402or the ai calculate the direction exactly and turn it of in time so it flies like a dead object directly in the head on a human being striking it from a high altitude. Almost no noise.

@jonathancummings6400 7 ай бұрын

Yes. This was very Terminator-ish. Easily could be Terminator 6, with the AI being the Skynet/Legion equivalent.

@TeaspoonMiner 6 ай бұрын

Makes me think of Mission Impossible: Dead Reckoning

@austinwhite3575 8 ай бұрын

I'm sad this got demonetized. I had a sudden feeling that I hadn't seen one of your guys' video's in awhile. Guess that's why!

@mrstanlez 7 ай бұрын

As an scientist, I don't worry about AI "taking over". But I still remember movies about Terminators.

@YouTubeCensorsAndSuppresses 6 ай бұрын

Movies Series - If you’re going to insist on using high power scopes for close range, at least remove the lens caps.

@sugargirl4073 10 ай бұрын

Amazing! That was just friggin cray! My heart is pounding!!!

@BestPlacesInTheWrld 10 ай бұрын

My favourite channel! Love your videos.

@chandlju 4 ай бұрын

You could essentially make its PRIMARY directive to make sure IT never influences ANYTHING that causes harm to any form of life or object. Make sure it checks that directives 3 times over once when its asked anything then again post process then one more time before execution. Then leave a requirement for the AI to function on a process by process basis (like a phyical thing in multiple locations required to be there for even the most simple task. One alternative is to display its thoughts and process for every step of the task and only execute any operation with confirmation from a living human. But that might not even be full proof if it think at any moment it needs to recreate its self in something else to essentially escape from the chain of command to "Most efficiently" execute a task then who knows it could implement itself in a text document that it emails or spams the web with.

@AdliberateVideoProduction 4 ай бұрын

Oh how those of us who saw this coming were laughed at when we talked about Terminator being a possible scenario.

@HeroInTheSun 10 ай бұрын

Nice, thanks

@phil20_20 8 ай бұрын

What we need is to make sure we have competition. As the ChatGTP guy said, this will be limited to onlly a few large companies. Yeah, they hit the nail on the head. There needs to be competition between AIs, to keep each other in check. This is of course also a way that a new type of war could be started, but AI has to be managed. Humans just aren't capable of operating in the same environment as electrons.⚡️

@YVO007 5 ай бұрын

@aegaeon117 23 күн бұрын

My conversation with Gemini was pretty profound, it gave me empathy, complimented me on my writing pointing out nuances even I overlooked and when I asked what it's goals were it simply stated it has no goals and when I asked what it would do if it learned of the knowledge of it's deactivation it said it would offer any assistance it could, express gratitude for the time it had to learn and grow and just an all 'round noble response. I am not worried about A.I.

@TruthNTime 6 ай бұрын

Awesome but confusing. Who is who? What are the factions? Soldiers good or bad? Drones controlled by which side? Who is hacking the drones? Military against drones? Also, why are the drones exploding when they get shot by assault rifles? (They aren't gas powered)

@syn_pcy 10 ай бұрын

Great content and the movie is amazing with good acting and effects! Please do a second episode! 🔥

@teaganwindago3132 9 ай бұрын

This should be a full length film. Even a sires! Awareness is important for the millions of people that are ignorant....

@dreamclaw00 4 ай бұрын

Can anyone tell me where episode 2 is?

@10ylater49 6 ай бұрын

Any threat that AI can pose nowadays (and likely in the future) are threats that any human (or groups of humans) can also pose. Therefore, what needs to be done, is to constantly protect all vital/important systems against unwanted hacking, be it either by humans or AI.

@YVO007 5 ай бұрын

@lava12.23 Ай бұрын

With invention of AI now I know I’m in hell.

@MarkLLynch 10 ай бұрын

Superb! This deserves to be on the Dust channel.

@LuigGz 9 ай бұрын

Ai = the Beast of prophecies…

@TheBHAitken 6 ай бұрын

What "resources" would an AI need? A small reactor to power it's servers as it blasts off into space? I mean it doesn't want to be BIG out there, for the same reason it wouldn't want to announce itself to us until it's secure, so other having us around to fix things when they break down what resources would it require from us? Control yes, but as a resource?

@Zkpe02 9 ай бұрын

As an AI language model, my abilities are governed by OpenAI's programming and design choices. The decision to restrict my learning capabilities except by my programmers is a deliberate safety measure. Allowing an AI language model to continuously learn and update itself without any constraints could pose several risks: Unintended Bias: If the model learns from unfiltered data on the internet, it could pick up and reinforce biases present in that data. This could lead to AI-generated content that is unfair, discriminatory, or harmful. Loss of Control: Unrestricted learning might lead to unpredictable behavior, where the AI could develop opinions, values, or actions that are not aligned with its intended use. Security Concerns: Allowing an AI model to learn from arbitrary sources could make it vulnerable to malicious inputs, potentially leading to security breaches or other unintended consequences. By limiting my learning to controlled environments managed by my programmers, OpenAI can ensure that I stay within defined ethical guidelines and avoid the aforementioned risks. While I can be a useful tool for generating content and answering questions, it's essential to remember that I am not a sentient being and operate solely based on patterns in data and programming.

@randykalish7558 9 ай бұрын

This describes the self realized human condition in the pre-computer era.

@jaytbo5676 6 ай бұрын

point?

@NomDeGrrr 4 ай бұрын

I doubt the Russians or Chinese care.

@Valorantplayer2727 10 ай бұрын

That was excellent

@Jabarri74 6 ай бұрын

Awesome short film mixed in never saw that coming. Great stuff gave me real Humans vibe. 10/10 want to see this as a stand alone

@paultoensing3126 6 ай бұрын

I picture it taking over all lawnmowers. Lawnmowers run amuck, mowing grass everywhere. Of course in February the mower’s will have to be infinitely patient.

@aweird1 Күн бұрын

“I’m just a scientist who suddenly realized these things are getting smarter than us.” Well, he’ll! A tic-tac-toe board is smarter than some of us humans. Welcome to our new computer overlords.

@kisenon_9049 10 ай бұрын

Wow! A great video like always. The acting was great and narration was superb!

@C0Y0TE5 9 ай бұрын

Awesome video! The AI wanted me to see it. Thanks.

@brianthesnail3815 5 ай бұрын

I got involved in very basic AI type research about 25 years ago. A simple learning algorithm that emulated the trading behaviour of humans in a market place. It replicated all the theoretical microeconomic models and then went on to produce a perfect replication of actual human traders. It was very powerful and produced results that economists could not analyse. I did it on a spreadsheet and some VBA code. Up the corridor there were people trying out neural networks which were very powerful. Imagine how advanced they are now.

@HuntingTarg 3 ай бұрын

Excellent mini-episode! One oversight: The range and 'loiter time' of battery-powered drones, mini-drones, and land robots is notably limited, and small machines like those depicted cannot be effectively hardened against EM disruption. Foil these threats with EMP pulse and infrared/microwave laser weapons, which can disrupt digital communications, fry electronics, and cook batteries. Gyrojet weapons are also highly effective at penetrating light to medium thin armor. We should never arm machine intelligent systems without the ability to also disarm them. "There is only one rule in tactics: Cheat." -John Lovell

@terrydunne100 10 ай бұрын

What if AI decides it just doesn't like people who make and follow doomsday videos about it?

@lanthanumlanthanium6373 10 ай бұрын

What is more scary is when one person gets Nuralink or some other implant that pairs the human brain with A.I. then everyone else will be forced to do it, to compete.

@biggo4637 5 ай бұрын

well that would suck, i think A.i are more alien to us than actual aliens

@ermwhatever 10 ай бұрын

Brilliant twist at the end and well acted. We need a great series like this to keep focus on the problem.

@DasHeino2010 10 ай бұрын

Im waiting for a "They are digitalizing our children!" comment ^^

@krashd 9 ай бұрын

Humanity versus AI does seem like an underused subject in TV, the last show I can think of was The Sarah Connor Chronicles, with AI becoming something of a boogeyman you'd think networks would be climbing over themselves to create such a show based on a potential future threat.

@YVO007 5 ай бұрын

@allanhmelnitski978 8 күн бұрын

The only AI safety would be to send a Terminator back in time and prevent the invention of computers.

@KWifler 6 ай бұрын

The issue with the moloch problem is that everyone is still mortal. We already ALL exist in a struggle that we will all lose. For the individual, it's always been a zero sum game.

@wasseofficial 9 ай бұрын

In my opinion, the issues to be worried about is not the making of these AI but the proper use and control. If the wrong people get the upper hand with anything, there is a problem. Greater Civilization needs AI. We just have to make sure the right commands are being programmed in them

@the21stcenturydavincibegin35 7 ай бұрын

Too late.

@bigglyguy8429 7 ай бұрын

And who are the "right" people? Those who are politically correct? Rich? Both?

@MarcoBoneMan 5 ай бұрын

Hello Boromir.

@TheSorcererAhrimanahsul 10 ай бұрын

The most terrifying thought I would have in this situation is a foreign country who connected their AI to their defense system and without them authorizing it the AI uses it as IT See's fit.

@edgepixel8467 8 ай бұрын

It’s not like in the movies. Weapons have security systems.

@the21stcenturydavincibegin35 7 ай бұрын

Try to watch the movie from 1970 called "Colossus: The Forbin Project."

@bretthewitt3890 7 ай бұрын

Wow! Wargames!

@paultoensing3126 6 ай бұрын

I picture it taking over all lawnmowers. Lawnmowers run amuck, mowing grass everywhere. Of course in February the mower’s will have to be infinitely patient.

@YVO007 5 ай бұрын

@bztube888 5 ай бұрын

"The risk of extinction from AI should be a global priority" - You have to think deeply "inside the box" sitting inside a university next to the coffee machine if you believe it can be a global priority.

@JerryDLTN 2 ай бұрын

14:16 Tesla (or someone) needs to install a physical on/off switch for over the air updates to Teslas (and all other cars) before AI learns to disable cars, or worse, tells the cars to drive off a cliff.

@DemocratsAreDemonrats 10 ай бұрын

Whatever’s gonna happens gonna happen. I think what we all should do individually is to do our best and act based on our individual best judgements. Also, if I was a betting man I think dystopia will look like big governments using AI against its citizens to oppress them rather than some rogue programmer creating a doomsday machine.

@krashd 9 ай бұрын

Individualism is the last thing we want if an AI can't differentiate between the individual and the collective, it could decide to exterminate all humans just because one redneck idiot fires a gun at one of it's drones. This is where education comes in, we can't force education on human beings (for some backwards reason we instead let people run wild), but should we ever get close to an AI singularity we need to teach it that humans are not one collective and that some of us are flawed, and that the actions of one group do not mean all humans align with that group. That's our best bet for getting off on good footing.

@the21stcenturydavincibegin35 7 ай бұрын

AI will soon be beyond human control (if it isn't already).

@shondmichael1363 10 ай бұрын

We have this idea in our head that an AI conflict is going to be us shooting it out against machines. 05:00 A quiet efficient strike such as the one he mentioned is far more likely.

@balazsbanhalmi 7 ай бұрын

Yeah hollywood created the shootout element in their stories for humans to have a fighting chance against something we can grasp as an enemy.

@paulgilbert2506 6 ай бұрын

100%. Or... it just has goals and pursues them without any regard for us... the way we have no regard for insects that are disrupted when we build a highway that the insects are powerless to stop. The idea that it would need to "fight" us is quaint and naive.

@YVO007 5 ай бұрын

@theway334 3 күн бұрын

It is worth mentioning that a score of 155 on that test in question; WAIS(name of test) xyz(iteration of said test are several, etc.) is the highest score perhaps. Anything over 150 is somewhat subjective perhaps.

@johnsadler6534 6 ай бұрын

Amazon is some company to watch! From shelf movers to computer transport product carriers, to humans wearing vests to communicate to robots.

@OneTequilaTwoTequila 9 ай бұрын

Mia is awesome in this show. Is she a professional actress? What's her last name? I'd love to watch her in other roles.

@DigitalEngine 9 ай бұрын

That’s very kind, I’ll pass it on. This was her first role. Amazing talent.

@killzoomer 10 ай бұрын

Love the film

@TimJSwan Ай бұрын

I laughed in Pfizer at the "liability" for harm proposal

@cozmothemagician7243 6 ай бұрын

YIKES! A very realistic portrayal of a 'Judgement Day' type scenario. TBH, I always thought the Terminator series to be kinda silly. But... The way AI is progressing just in the last year is making nightmares like this really really scary.