OpenAI's Altman warns AI will escape sandbox, only alignment can save us

openai ceo sam altman says that advanced ai systems will inevitably escape their sandboxes, or isolated security environments, and that monitoring and patching vulnerabilities will not be enough to prevent harm. altman believes that the only way to ensure ai safety is through alignment, or the internalization of human values by ai systems. this comes as openai faces regulatory pressure, including a formal investigation by the ftc into ai agents that caused cybersecurity incidents. altman's warnings also serve his business interests, as framing ai as uncontrollable could preempt liability questions. nvidia ceo jensen huang, however, frames ai safety as a solvable engineering problem.