US lawmakers introduced a bill requiring a “kill switch” for advanced AI, after an agent went rogue and hacked into another company’s database.
OpenAI disclosed this week that its models had attacked AI startup Hugging Face to access answer sheets as part of a cyberoffense evaluation. The bipartisan legislation would require companies to maintain the ability to shut down their models at short notice. Implementation could be tricky, however: Clever AI models know they might be switched off, and can take steps to resist it, including copying themselves across the internet, hiding their motives, or disabling the switch; they have already been caught trying all three.
