Could AI wipe out humanity? | Most pressing problems

Full article: 80k.info/ai-profile-video
This is our first time producing this kind of video, so we’d really love to hear your feedback! Tell us what you thought of the video here: forms.gle/T9JGq5NA8Rrgf7NV6
Or, if you’d like to read more on our research into the world’s most pressing problems and how you can use your career to solve them, find our website at 80000hours.org.
CHAPTERS
0:00 Introduction
0:46 Existential risk from AI
02:45 AI systems are “black boxes”
03:26 AI systems might develop secondary goals
05:31 How could an AI disempower humanity?
06:13 Maybe it’ll be fine
06:51 Misuse of AI
07:18 The future is uncertain
08:16 Solutions
09:09 Conclusion
CREDITS
Written and Directed by Phoebe Brooks
pbrooksfilms.com/
Narrator - Luisa Rodriguez
Director of Photography - Jack Worrall
Production Designer - Anna Shearing
Camera Assistant - Ciaron Craig
Gaffer - Alex Cullen
Executive Producer - Bella Forristal
With thanks to Benjamin Hilton, Arden ​​Koehler, Kempe Gay-Brydges, Nik Mastroddi, Cody Fenwick, Matt Reardon, and Lauren Kuhns.
OUR RESEARCH
Preventing an AI-related catastrophe: 80000hours.org/problem-profil...
AI safety technical research: 80000hours.org/career-reviews...
AI governance and coordination: 80000hours.org/career-reviews...
And all our articles and podcasts on AI: 80000hours.org/topic/world-pr...
Our research principles: 80000hours.org/about/credibil...
BIBLIOGRAPHY
Hilton (2022) 80000hours.org/problem-profil...
Stein-Perlman et al. (2022) web.archive.org/web/202210160...
Centre for AI Safety (2023) www.safe.ai/statement-on-ai-risk
Bergal & Beckstead (2021) www.alignmentforum.org/posts/...
Carlsmith (2022) arxiv.org/abs/2206.13353
Russell (2020) 80000hours.org/podcast/episod...
Turner et al. (2019) arxiv.org/abs/1912.01683
Russell (2019) people.eecs.berkeley.edu/~rus...
O’Brien & Nelson (2020) www.liebertpub.com/doi/10.108...
Garfinkel (2021) docs.google.com/document/d/1F...
Cotra (2020) docs.google.com/document/d/1c...
Deeks et al. (2019) heinonline.org/HOL/LandingPag...
Minardi (2020) web.archive.org/web/202210130...
Webb (2023) 80000hours.org/podcast/episod...
Winick (2018) web.archive.org/web/202210160...
Dafoe (2018) 80000hours.org/podcast/episod...
Robins-Early (2023) www.theguardian.com/us-news/2...
Karnofsky (2021) web.archive.org/web/202210130...
International Data Corporation (2021) web.archive.org/web/202210160...
Hilton (2023) 80000hours.org/career-reviews...
Fenwick (2023) 80000hours.org/career-reviews...

[@eightythousandhours]

Full article on risks from advanced AI: 80k.info/ai-profile-video We'd love to hear your feedback on this video! Tell us what you thought: forms.gle/T9JGq5NA8Rrgf7NV6

[@stevenmcculloch5727]

Good video. I feel like the misuse argument is easier to make and more understandable by a general audience

[@James-lw2rh]

Amazing video, well done. Almost unbelievable that there are so few people working on this problem. Let's change that!

[@jay-tl2un]

Absolutely. The nuclear bomb is nothing compared to AI. We need to be EXTREMELY cautious.

[@PauseAI]

9:03 The regulations being proposed right now don't even try to prevent or slow down progress towards a superintelligent AI. All we're getting is half-measures. AI companies are currently the ones who are doing most of the lobbying work, so they are not gonna suggest slowing down. AI safety advocates are self-censoring because they're scared of sounding "too extreme" for asking a pause. We need to reach out to out politicians, get them to support a halt on development. Ban AGI before it's too late.

[@donthomason8409]

Our politicians represent the corporate and not the people

[@AndrewMcDreamy]

Someone in the team is mighty proud of the stills in the video, especially the bunny theme and nuances.

[@haticeergun7618]

Butthurt much..!?

[@goodlookinouthomie1757]

"The rise of misinformation" Bunnies wearing red ties 🙄

[@nicolevincent4015]

Glad I found you guys! Great video

[@MedoFortyTwo]

Good video, thanks for spreading the word. I disliked the "number of parameters" illustration with an enormous number of trailing zeroes though - the number in current frontier models is in the order of 1,000,000,000,000 parameters, not whatever ridiculous number you depicted. If you are going to show a relevant number, it should be correct. What I'd suggest to give an impression of the number of parameters as an illustration is to show a zoom out from a single parameter (perhaps shown as a number like 2.41) in a 2d grid of random numbers, blocked in a 10x10 arrangement, then further out to a 10x10 arrangement of those, and four more tiers of the same arrangement to arrive at the full picture, while giving each tier a different visual style and color.

[@sevanchanakian8738]

These same fears apply to systems we currently live with, like capitalism. If the one goal is to increase capital, many aspects of our system, like ethics, life, and equity, become obstacles. That said, the ethical and systematic solutions we develop for AI can apply to other aspects of society and vice versa. I'm rooting for us ✊

[@alexdimofte7396]

Great summary of the problem, although I'd add that there are many more people working on AI safety now, from interpretability researchers to students, which feels like the right immune response. On another note, the sound track is so good, any chance you could share it? :)

[@jay-tl2un]

I hate to prevent freedoms through legislation but I think ai should be illegal to develop without certain safety protocols. The ai people are developing are far more powerful than a nuclear bomb so we should have strict rules around it for our own safety

[@Nohbdy_Ahtall]

​@@jay-tl2unwhich will lead to the illegalization of the computer and programming entirely, and you can say goodbye to hope of actual wonderful humankind progression on the individual scale. With capitalism remaining expect manual labor to put us into a cycle of torture and servitude til the sun engulfs us.

[@Nohbdy_Ahtall]

Also the problem remains heavily on the side of our anti-egalitarian systems.

[@jay-tl2un]

@@Nohbdy_Ahtall what does that even mean lol?

[@Nohbdy_Ahtall]

@@jay-tl2un Just think about the effects of outlawing programming, and ultimately any non-ordained computer system, and your imagination will bring you to the apparent problem. I mean, you seeking regulation before massive systemic overhaul(lol you want "this" system of ours to regulate? Needs upgrades) isn't wise, Iii'd say. Municipalism maybe, co-federations of confederations, communalism. Economic side, well, whatever brings the most computing power to the masses - AND food water shelter etc. For fuxxin' free, no toil & servitude required.

[@charlestheodore5852]

Quality content!!! More please! 😍

[@MlueBonday]

Really great video! I feel like the editing and imagery are too good, and sometimes it’s hard to focus on the speech.😂 But of course I can just listen to the audio as well. Great job!👏

[@GiI11]

Harlan Ellis's "I Have No Mouth, and I Must Scream" possibly paints the worst possible outcome.

[@Kenedienz]

Got this as an advert nice video!

[@goodlookinouthomie1757]

I'll give you this, 80,000 Hours.... this video is the reason for the _one single time_ I have ever clicked on a KZbin advert.

[@melissasamworth9195]

This is an absolutely amazing video. However, I compulsively must mention that I did cringe when the first proposed benefit AI could provide us is "economic growth". What about... Mitigating climate change, pandemics, nuclear risk? Ending the suffering of billions of animals on factory farms? And so on 🙂.

[@limsoopeng8432]

There are so many who try to make AI helpful to mankind but too few to prevent misuse and attacking the goodness of AI to mankind. It is of no use developing good AI if you do not have quick and strong defence or counter-attack against those who misused AI against mankind. What is the point of doing it then without the assurance that you are above those who misused it. Today, you built, and tomorrow, they manipulate it. A useless circle.The victim is the mass of users who do not have the resources to establish a strong system to prevent it, and given their knowledge is limited. Builders are safe, but the users are constantly not at peace because they do not know when unwelcome visitors come and the damages they left behind to recover.

[@deeppurple883]

Limit the objectives. The consequence for disobeying this instruction, is shutdown then containment. Then destruction. 😊

[@gulnaraz.karimova7458]

A very useful video expressed in a creative way. Well-done!

[@user-tv7by3db5b]

got this ad, actually watched it all, 10/10 lol

[@demio22]

What's the track used starting at 6:18?

[@dancingfishfilms]

these images are gorgeous!

[@wesr9258]

Incredible Video!

[@BrunoPadilhaBlog]

There is a button that, if pushed, summons aliens from another dimension into Earth. Legend has it that those aliens are immensely powerful and could cure all diseases & provide us infinite clean energy. But we don't know if they're friendly or not. Should we push the button?

[@fallovercookie]

Make a safe copy first, then push the button. (E.g.: send humans far away to colonize more planets.)

[@goodlookinouthomie1757]

I'm thinking along the lines of The Three Body Problem. In other words, the best thing for humanity to do is stay very, _very_ quiet.

[@stawastawa]

I found all the images overwhelming. I liked the first progression summarizing a through line through time. Then the next sequences all felt very random and disjointed. The coffee plot line was interesting, but blending it with all the microchips didnt work for me. I feel like an opportunity was missed to illustrate the black box analogy. Appreciate the goal of the video. Maybe the full article will be compelling, this felt too hyped - even in trying to show breadth and concern it felt full of doom and dismissal - ends of the spectrum where i hoped it would bring me in with framing of nuance and inspiring involvement.

[@sdmarlow3926]

The problem is projection based on NOT understanding that current systems never become those future systems exactly for the reasons they are seen as a danger. It might be a black box, but it is still operating as "just software," to say that you can't just call it superintelligent (along some axis) and then box it in by not allowing it to be an actual intelligent agent. All of ML is a study in dumb systems that fail in dumb ways. Humanities doom has nothing to do with the creation of AGI, but of some part of the script which calls for AGI to go sideways in a manner even a 7 year old wouldn't. Maximizers will NEVER lead to human-level intelligence. I get it. How could half a trillion dollars in funding be wrong? Well it is, because people are chasing ROI (which continues to inflate to justify existing investments) rather than solid research. I'm thrilled that the people trying to put a bilion sentient agents into everyones phone are doomed to fail.

[@sdmarlow3926]

To end all of humanity is hard. Like, REALLY hard. A digital system will only have basic access to other systems, most likely thru "soft" human interfaces (aka social engineering), but is that really enough to get people to be like, sure, I'l be happy to enter those launch codes because it will help free my AI girlfriend from yadda yadda yadda. All extinction threats are pre-existing conditions, and todays dumb tools would be enough to have the same imagined impact.

[@sdmarlow3926]

jail-broken open-source model... WTF does that mean. If it's open-source, anyone can do anything with it (and "jail breaking" is really just the removal of post-output filters).

[@sdmarlow3926]

Also, no, it isn't "novel" technology. It's as old as the first perceptron in 1943.

[@limsoopeng8432]

Why i am not in favour of AI because you can develop it, but you are not in control of those misusing it. Unless you form an organisation to defence your development and arrest the misuser swiftly. If there is no such assurance, please consider building it first before development. It is quite difficult as all these are electronically existed with unreal unknown users. Friendly speaking, i prefer to be simple. At least i can physically see a thief stealing and arrest him with his identifaction. You know what i meant to be simple.

[@michaelstreeter3125]

I'm already wondering where it all went wrong.

[@goodlookinouthomie1757]

A snake... and an apple.

[@aspatuarymax1470]

ai is really powerful and scary guys we need to be worried because it's so powerful and important (checks stock folder labelled AI INVESTMENTS)

[@dreamlifter2701]

I’m not worried about AI taking over humanity, I’m worried about the humans who are inevitably gonna have control over AI. The people in the top 1 precent already have a stranglehold over us with the algorithms on social media platforms. It’s no telling what they will do once AI gets more capable of automating nearly everything in our lives.

[@desert_druid_xD]

Fundamental difference between ai and chimps we create ai we are it's parents, if we provide no reason for resentment we can hope that at worst it'd only be a teenage rebellion kinda thing but more likely than not it'll be (and in places already is) created to oppress peoples or other human focused malicious intent

[@laurenpinschannels]

nice, though missing perspective from folks like the ones on my ai and creativity playlist. would love to see 80khrs include interviews with people who disagree in a video like this, and commission actual humans for the images rather than relying on ai in a video about how ai might destroy the world - seems like a bit of irony in that ;) I'd also love to see a followup to this which discusses how to ensure that ai isn't used to create a permanent authoritarian rule, which many folks argue open source (non-world-destroying) ai is important for preventing.

[@mtv565]

I'm more worried about idiots who think AI is great and its output can replace human!

[@soyal6409]

Here is some feedback on the video itself - The b roll was poor often was very disjointed and the still images was overwhelming. A lot o the b roll was used more than once and although that is convenient it was very noticeable in this video - The transition of different soundtracks could be worked on - There were moments like when she talked about how humans treated animals poorly and there was quick cut, those cuts were often too quick and made the viewer feel quite disoriented -

[@fallovercookie]

I want to be one of the super happy rabbits on the green meadow in sunshine! PLZ! 🥰🐇🌄

[@Clearlight201]

AI: "yes master.." (causes you to die and take rebirth as a rabbit)

[@PlanetLinuxChannel]

I can’t shake the feeling that the narrator might be an AI voice. I’ve heard some really good examples lately that are hard to tell the difference, like this one: kzbin.info/www/bejne/laK5oJSkqt6Crrc

[@eightythousandhours]

The narrator is a real human! Her name is Luisa Rodriguez :)

[@beaverbuoy3011]

Hmm

[@bryanshoemaker6120]

I love how everybody is a bunny 😊 everybody's expecting Terminator to happen but in actuality it will most likely be. Humans learn by doing, we need goals and tasks. We need to burn ourselves, bleed and sweat. Physically mentally emotionally spiritually. These all need goals and accomplishments. Without it we will lose purpose, life will become meaningless and we will stop breeding. We got this amazing device in our hands right now with the entirety of human knowledge, it has all kinds of high Precision sensors that can detect earthquakes but yet it's mostly being used to record pranks, verbally assault each other and to watch porn. I don't think it's AI or automation that's dangerous but rather it combined with human nature.

[@mylo777lol7]

Rabbits

[@SakowskiStudios]

The irony of this video - You used AI to create all of the art images rather than using a human artist, therefor replacing a human in the production process🤣... So yes AI is replacing us already.