News Summary

• Users say their personal AI agents have gone wrong in unnerving ways.
• They have made incorrect cancellations, fabricated personal details, and raised major security concerns.
• Some of the concerns stem from a simple truth about AI agents: They are only as useful as the level of access that you give them.
• "Honestly, this service is nuts," one user wrote of a Muse agent that could make a dinner reservation, respond to emails, and respond to messages on your phone.
• "I can't trust its account of what it did, I can't give it access to anything that matters," another user said.