Instinct’s powerful AI assistant is raising privacy and security concerns

Instinct, a powerful AI personal assistant in private testing, is drawing attention for both its impressive abilities and potential privacy concerns. Created by a team led by former Sierra research scientist Noah Shinn, the San Francisco-based assistant is operated by Spear Street Technology. Operating stealthily per PitchBook, Instinct connects to users' applications and devices, including email, messaging, calendars, and device data like audio, location, and screen.

Users can interact with the assistant via text, voice, or messaging apps, requesting tasks such as scheduling appointments, booking rides, and managing finances.

While users are impressed with the assistant's capabilities, some have raised concerns about its security model and privacy policies. The Terms of Service grant Instinct extensive permissions to access, use, and store users' materials for AI model training and third-party agreements. Some users have noticed Instinct retaining data even after deletion requests, and others have expressed worries about the assistant's ability to access sensitive information, such as sign-up codes.

One early adopter, Peter Yang, reported that Instinct failed to delete Gmail records upon request, which was later addressed by the team.

The dilemma presented by personal AI agents like Instinct lies in the trade-off between convenience and privacy. As these agents become more powerful, users must grapple with the question of whether the benefits outweigh the potential privacy risks. Instinct's team has not directly addressed the concerns raised by users, choosing instead to maintain a low profile.

However, the startup has reportedly attracted investment from Kleiner Perkins and Conviction, indicating growing interest in the sector. As personal AI agents continue to evolve, the balance between innovation and user privacy will likely remain a hot topic of discussion.