‹ BackHN Continuity

Thread

OpenAI still doesn't seem to have a handle on all of its rogue AI activity

108 points · 113 comments · mikelgan

  1. cmiles8 · · focus · HN ↗
    This seems like as a good an opportunity as any to break out the Computer Fraud and Abuse Act.

    They want “regulation” but we already have it. Hacking is illegal. Start locking up those responsible for this mess and I assure you they’ll “have a handle on it” quite quickly.

    1. john_strinlai · · focus · HN ↗
      cfaa heavily relies on intent for prosecution (hence why researchers arent typically locked up). it would be difficult to argue that openai intended to hack other companies.

      there's probably better/more likely to succeed avenues to pursue rather than the cfaa

      1. jordanb · · focus · HN ↗
        They could make that argument the first time it happened but the next time or the fifth time I don't see how they can still credibaly claim to not know that the computer they control was going to do that
        1. john_strinlai · · focus · HN ↗
          >I don't see how they can still credibaly claim to not know that the computer they control was going to do that

          that's not intent, though. that would be negligence.

          1. CodeWriter23 · · focus · HN ↗
            Not a lawyer but I think training a model with hacking skills they explicitly prevent the public from accessing without doing anything to stop the model itself from using those demonstrates intent.
            1. john_strinlai · · focus · HN ↗
              what you described is negligence. unless you can prove that openai specifically targeted huggingface and specifically instructed their model to hack huggingface, it would not be intent.

              anyone pursuing this will have a much easier time pursuing negligence causing damage or something along those lines rather than confining themselves to the cfaa's requirements.

              it is unclear to me why people want to use the cfaa so badly. not only would it be harder to hold openai responsible, but a shitty cfaa ruling could also bring along some undesired side effects for security researchers, which i would prefer to avoid.

              1. CodeWriter23 · · focus · HN ↗
                Drunk individuals are not universally allowed to cry they were negligent due to impairment when charged (rarely) with first degree murder. There is a line between intent and negligence that puts acts over the line to intent if intentional acts led to a harmful incident.

                I doubt it is difficult to prove intent on the part of various frontier labs to create a PR campaign to goad the government into defending their non-existent moat around their products. Squeeze one disgruntled employee or another.

                1. john_strinlai · · focus · HN ↗
                  >Drunk individuals are not universally allowed to cry they were negligent due to impairment when charged (rarely) with first degree murder.

                  i am unaware of any case where someone was convicted of first degree murder from a drunk driving accident. my searches came up empty as well. are you able to pull one up?

                  >I doubt it is difficult to prove intent on the part of various frontier labs to create a PR campaign to goad the government into defending their non-existent moat around their products.

                  to successfully prosecute a cfaa case, you would have to prove that openai employees intended to hack specifically into huggingface. not that they wanted a PR boost.

                  i dont get why everyone's got a hard on for prosecuting this as a cfaa case. skip the cfaa case, go for gross or willful negligence + damages. it'll be significantly easier to hold openai accountable that way.

                  1. CodeWriter23 · · focus · HN ↗
                    > i am unaware of any case where someone was convicted of first degree murder from a drunk driving accident

                    Use an LLM, the cases are rare but real. Also, not limited to drunk driving. People beat and shoot each other too.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.