Close Menu
    Zukunftsforschung

    OpenAI-Insider schätzt, dass die Wahrscheinlichkeit, dass KI die Menschheit vernichten oder ihr katastrophalen Schaden zufügen wird, bei 70 Prozent liegt

    Share.
    View 24 Comments

    24 Kommentare

    1. Maxie445 on

      „In an [interview with *The New York Times*](https://www.nytimes.com/2024/06/04/technology/openai-culture-whistleblowers.html), former OpenAI governance researcher [Daniel Kokotajlo](https://futurism.com/openai-safety-worker-quit-confidence-agi) accused the company of ignoring the monumental risks posed by artificial general intelligence (AGI) because its decision-makers are so enthralled with its possibilities.

      „OpenAI is really excited about building AGI,“ Kokotajlo said, „and they are recklessly racing to be the first there.“

      Kokotajlo’s spiciest claim to the newspaper, though, was that the chance AI will wreck humanity is around 70 percent — odds you wouldn’t accept for any major life event, but that OpenAI and its ilk are barreling ahead with anyway.“

      The term „[p(doom)](https://www.nytimes.com/2023/12/06/business/dealbook/silicon-valley-artificial-intelligence.html),“ which is AI-speak for the probability that AI will usher in doom for humankind, is the subject of constant controversy in the machine learning world.

      The 31-year-old Kokotajlo told the *NYT* that after he joined OpenAI in 2022 and was asked to forecast the technology’s progress, he became convinced not only that the industry would achieve AGI by the year 2027, but that there was a great probability that it would catastrophically harm or even destroy humanity.

      As noted in the [open letter](https://righttowarn.ai/), Kokotajlo and his comrades — which includes former and current employees at Google DeepMind and Anthropic, as well as Geoffrey Hinton, the so-called „Godfather of AI“ who [left Google last year](https://futurism.com/the-byte/godfather-ai-quits-google) over similar concerns — are asserting their „right to warn“ the public about the risks posed by AI.

      Kokotajlo became so convinced that AI posed massive risks to humanity that eventually, he personally urged OpenAI CEO Sam Altman that the company needed to „pivot to safety“ and spend more time implementing guardrails to reign in the technology rather than continue making it smarter.

      Altman, per the former employee’s recounting, seemed to agree with him at the time, but over time it just felt like lip service.

      Fed up, Kokotajlo quit the firm in April, telling his team in an email that he had „lost confidence that OpenAI will behave responsibly“ as it continues trying to build near-human-level AI.

      „The world isn’t ready, and we aren’t ready,“ he wrote in his email, which was shared with the *NYT*. „And I’m concerned we are rushing forward regardless and rationalizing our actions.“

    3. KhanumBallZ on

      The only safeguard is open sourcing and decentralization.

      Don’t spend a penny on AI services. Freeload shamelessly and use locally run whenever possible

    4. Extreme-Edge-9843 on

      What’s that saying about insert percent here of all statistics are made up?

    5. revel911 on

      Well, there is about a 98% chance humanity will fuck up humanity …. So that’s better odds.

    6. sarvaga on

      His “spiciest” claim? That AI has a 70% chance of destroying humanity is a spicy claim? Wth am I reading and what happened to journalism?

    7. Yamochao on

      Ah, yes, measured by all of the data points we have on other species, 70% of which were observably destroyed and harmed by AI

    9. higaki_rinne on

      If you know the kind of people that would work there then this shouldn’t surprise you. These people have a unique type of brain damage.

    10. michael-65536 on

      Anyone who thinks a machine will be more savage than human beings doesn’t know human beings.

      To our knowledge there has never been a creature on earth as wantonly and gleefully vicious as homo sapiens. (Not even cats, though they do try.)

      Speculating that something not specifically adapted for savagery could compete with us on those grounds by mere accident is a joke. (Or a premeditated rhetorical tactic to garner media attention for leveraging the ludd dollar.)

    12. rrfe on

      I would take this with a pinch of salt. In 2008-2011, social media was going to save humanity because of the election of Obama and the Arab Spring. When Trump used those same tactics in 2016, it became an existential threat.

      The early prognosticators were wrong about social media, and it’s just as likely that they would be wrong about AI.

    13. Lord_Vesuvius2020 on

      I’m sure “70%” was given but as others have commented it’s not clear what that even means. Based on what? And the idea that open source is some kind of protection seems totally bogus. We all know the huge amount of data, the huge computing resource, the huge power requirement just to be in this game. You need billions of dollars to do this (or else be a government with similar assets). I am still finding that AI chatbots make mistakes. I asked Gemini yesterday (June 8) when the new episodes of “Bridgerton” were being released and it told me that these episodes were already released and this happened on June 13. I think there’s a way to go before we get to “singularity” with these guys.

    14. IAmWeary on

      It’s not AI that will destroy humanity, at least not really. It’ll be humanity’s own shortsighted and underhanded use of AI that’ll do it.

    15. gza_liquidswords on

      Might as well say „people that watched Terminator estimate 70 percent chance that AI will destroy or catastrophically harm humanity“. This AI hype is so dumb, in its current form it is Clippy with more computational power.

    16. _CMDR_ on

      We as a civilization must stop the ruling class from developing autonomous murder robots or they will be able to end liberty for hundreds of years.

    17. Divallo on

      „The world isn’t ready, and we aren’t ready,“ he wrote in his email, which was shared with the *NYT*. „And I’m concerned we are rushing forward regardless and rationalizing our actions.“

      Pretty much. If it makes you feel any better humanity couldn’t show restraint even if they wanted to.

    18. givin_u_the_high_hat on

      Nvidia has said every country should have their own sovereign Ai. So what happens when these Ais are forced to believe cultural and religious absolutes? What happens when the Ais are programmed to believe people from other cultures deserve death? And what happens when they get plugged into their country’s defense network…

      [https://blogs.nvidia.com/blog/world-governments-summit/](https://blogs.nvidia.com/blog/world-governments-summit/)

    19. zach19314 on

      How many doomsday events are we going to have to worry about really? Humanity had been harming humanity since humanity existed. We will just have to figure it out like we always do once the problems arise.

    20. PartyClock on

      Probably but their fancy fucking word calculator isn’t going to be the thing to do it

    21. ChrysMYO on

      His assertion that they may reach AGI is the biggest leap of a claim he’s making. But I have no doubt these giant corporations aren’t doing enough to keep us safe from AI’s impact on society.

    22. presentaneous on

      Anyone that claims generative AI/LLMs will lead to AGI is certifiably deluded. It’s an impressive technology that certainly has its applications, but it’s ultimately just fancy autocorrect. It’s not intelligent and never will be—we’re literally built to recognize intelligence/anthropomorphize where there is nothing.

      No, it’s not going to destroy us. It’s not going to take everyone’s jobs. It’s not going to become sentient. Ever. It’s just not what it’s built to do/be.

    Leave A Reply