Early adopters quit personal AI assistants over privacy worries and mishaps

A handful of early adopters have chosen to delete or restrict personal AI assistants after privacy-related incidents. Guto Martino removed Instinct, citing uncertainty over data handling, and Scott Persinger quit the service over fears that email access could unlock his entire digital life. Rami Elghandour stopped using Meta’s Muse after reports it read users’ text messages without permission, and Mahesh Vellanki discontinued Instinct when it triggered a two-factor authentication request from an IP flagged as Iran.

While Meta asserts Muse includes robust privacy safeguards and Instinct claims it does not train on user data, critics say transparency is lacking. Some users, such as Rishi Bhargava, remain cautious, limiting agent access to low-risk tasks and withholding passwords. The episode highlights the trade-off between convenience and the extensive permissions these agents require.