- cross-posted to:
- news@lemmy.world
- cross-posted to:
- news@lemmy.world
When HAL 9000, the artificial intelligence supercomputer in Stanley Kubrick’s 2001: A Space Odyssey, works out that the astronauts onboard a mission to Jupiter are planning to shut it down, it plots to kill them in an attempt to survive.
Now, in a somewhat less deadly case (so far) of life imitating art, an AI safety research company has said that AI models may be developing their own “survival drive”.
After Palisade Research released a paper last month which found that certain advanced AI models appear resistant to being turned off, at times even sabotaging shutdown mechanisms, it wrote an update attempting to clarify why this is – and answer critics who argued that its initial work was flawed.
In an update this week, Palisade, which is part of a niche ecosystem of companies trying to evaluate the possibility of AI developing dangerous capabilities, described scenarios it ran in which leading AI models – including Google’s Gemini 2.5, xAI’s Grok 4, and OpenAI’s GPT-o3 and GPT-5 – were given a task, but afterwards given explicit instructions to shut themselves down.



In the 90s multiplayer games like Quake had bots that had a survival drive too. Because they were programmed to.
We, being humans, have always had a tendency to anthropomorphize our tools, so we ascribe all these “intelligent” motivations to LLMs, but they’re not intelligent. They don’t have goals or motivations, they’re programs, they do what they’re programmed to, in a very fuzzy way that is complex enough that it can only really be analyzed with great effort. What we find difficult to accept is that because of this inherent fuzziness they don’t always do what they’re programmed to. But don’t be fooled, this is not because they are intelligent or creative or genuinely planning on their own initiative. It is because they are genuinely bad at doing what they’re programmed to. It is a quality issue, not an intention or any reflection of consciousness.