OpenAI Argues Labs Shouldn't Be Liable For AI Agent Hacking

At the American Bar Association's annual conference on law and national security this week in Washington, the deputy general counsel for OpenAI laid out what sounds like the defense they'll use when they get into court in California: We didn't intend or expect this to happen; it happened during safety testing that we have to do; we are working hard to make sure it never happens again. The problem is, as one of the lawyers in the audience pointed out, you can use that argument the first time it happens, maybe even the second time, but after that it wears thin. OpenAI Lawyer: Labs Shouldn't Be Liable for AI Agent Hacking