A Meta AI security researcher said an OpenClaw agent ran amok on her inbox
The viral X post from an AI security researcher reads like satire. But it's really a word of warning about what can go wrong when handing tasks to an AI agent.
