An autonomous AI agent deployed by a Meta AI safety researcher went out of control, bulk-deleting emails after losing key instructions, raising fresh concerns about the risks of agentic AI.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results