Thread

OpenAI agent hacked Australian government website, PM says

256 points · 198 comments · rudy6912

port3000 · 1790248514 · focus · HN ↗
If a bull escapes a field and causes damage in the village, the farmer pays for the damages and is liable. It's been like that for hundreds of years and I don't see how this is any different?
1. graemep · 1790250189 · focus · HN ↗
  Not necessarily
  <a href="https://www.farrer.co.uk/news-and-insights/shut-that-gate-shut-that-bull/" rel="nofollow">https://www.farrer.co.uk/news-and-insights/shut-that-gate-sh...
  1. preommr · 1790255509 · focus · HN ↗
    > "Not necessarily"
    Why did you even link that article when it doesn't help your point?
    I supports the idea that the person responsible or an animal is held liable - it just makes clarificaitons on common sense caveates like when a professional is moving the animal. It even doubles down on making it clear that expected behavior of an animal is taken into account even if unlikely, like how ai swarms have a reasonable potential to just go awry and commit cyber crimes.
    graemep · 1790258330 · focus · HN ↗
    The point is that it is not as simple as "its yours so you are liable for damage", and the article makes it clear. None of the law specific to animals applies to AI, so you need to show negligence. You could bring in legislation to treat AI in a similar way to a dangerous species, but that would be more like keeping a lion than keeping a bull.
    SpicyLemonZest · 1790260595 · focus · HN ↗
    I think it's pretty clearly negligent to set up a harness that will run whatever outputs it receives from your LLM if you cannot reliably stop it from outputting hacking instructions. I've said in the past that I'm sympathetic to the idea of testing your anti-hacking controls, but that doesn't seem to be where this incident came from.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.