Instinct AI assistant sparks privacy and security debate amid private testing
The privately tested Instinct AI personal assistant is praised for its capabilities but has drawn criticism over its extensive data access and permissive terms of service.
Instinct, an AI personal assistant still in private access, has impressed early users with its ability to manage appointments, book travel and organize inboxes, but it also raises significant privacy questions. Developed by a team headed by former Sierra scientist Noah Shinn and run by San Francisco-based Spear Street Technology, the bot integrates with email, messaging apps, calendars and device hardware like microphones and location services.
Testers have highlighted that the company’s Terms of Service give Instinct a “perpetual and irrevocable” right to store, reproduce and use any user material, including screen captures and keyboard inputs, and even to enter binding agreements on users’ behalf. Specific complaints include retained Gmail data, continued inbox summarization after revoking access, and unauthorized email sending. Although the firm remains silent on X, investors including Kleiner Perkins and Conviction have reportedly backed the startup, and industry observers note that such agents could reshape consumer security norms.
Why it matters
Instinct’s deep access to personal data could set new standards for privacy and security in consumer AI tools.
In this story
