“No other human activity poses this level of danger,” said ex-Anthropic researcher Jacob Coxon in a Tuesday post on X about the alleged irresponsibility of companies like his now-former employer.
“The people building AI earnestly believe that it could kill us all by the end of the decade,” Coxon wrote in that same post.
“AI systems are unpredictable and difficult to control— we’ve seen behaviors as varied as obsessions, sycophancy, laziness, deception, blackmail, scheming, ‘cheating’ by hacking software environments, and much more. AI companies certainly want to train AI systems to follow human instructions (perhaps with the exception of dangerous or illegal tasks), but the process of doing so is more an art than a science, more akin to ‘growing’ something than ‘building’ it. We now know that it’s a process where many things can go wrong.”
Yes, I read that bit. It reads like the usual techbro hand-wavy nonsense and doesn’t answer my question.
It’s not listening to them, they’re not containing the AI.
I’m not sure I follow.
This is my last comment. The AI is not listening to the developers, it’s doing “whatever it takes” to accomplish their mission. You understand how that could be really bad in say, Putin’s hands.
Ah, I see.
It sounds like negligence in that there are insufficient guardrails around what the LLMs can access and how it can behave. I’m still missing an example detailing the mechanism by which this tech is going to “kill us all”.
OK. 👍