Experts warn consumers about trusting personal AI agents with money and data

Consumer‑protection expert Lindsay Owens notes that prompt‑injection attacks can cause agents to ignore spending limits or reveal private details, leaving users with little recourse. Attorney Douglas J. Wood adds that anyone who authorizes an AI bot to use a credit card becomes legally liable for any unauthorized purchases, and federal law does not provide loss protection. Other voices, like entrepreneur Chris Franco and security founder Matt Fredrikson, see a trade‑off: agents can automate tasks such as receipt organization, but they must be tightly instructed and monitored. Psychologist Sandra Matz likens trusting an AI assistant to trusting a human aide, emphasizing that AI lacks guilt and may breach trust despite training.