Instinct, an AI personal assistant currently in private testing, has garnered attention for both its impressive capabilities and potential privacy concerns. Developed by a team led by former Sierra research scientist Noah Shinn and operated by Spear Street Technology, Instinct connects to users’ applications and devices, including email, messaging apps, calendar, and device audio, location, and screen data. Users can interact with the agent via text or WhatsApp to perform tasks such as booking appointments, scheduling rides, organizing information, and handling shopping. Despite being praised for outperforming expectations, Instinct has raised concerns among testers regarding its security model and terms of service.

The company’s Terms of Service grant Instinct a “perpetual and irrevocable” license to access, use, and modify users’ materials, including for training its AI models. The terms also allow Instinct to receive detailed information from users’ devices, such as screen captures and keyboard inputs, and to enter into binding agreements on users’ behalf. Testers have reported issues such as Instinct failing to delete Gmail records upon request, continuing to summarize inboxes after access disconnection, and storing emails in plain text. Concerns have also been raised about the ease with which Instinct can be phished and instances where the agent sent emails on users’ behalf without prior consent. These issues highlight the trade-offs between the convenience of hyper-personalized AI tools and the potential loss of privacy and control. As of the article’s publication, Instinct’s team had not publicly addressed these concerns. Source

This is a summary. Read the original at techcrunch.com.