Instinct’s AI Assistant Is Drawing Praise — and Privacy Questions

TechCrunch reports that Instinct, an AI personal assistant still in private access, is drawing attention for both its capabilities and its privacy implications. Early testers have praised the product, while others have raised concerns about its security model and terms of service. Because Instinct remains in private testing, those concerns have not yet scaled to a wider public user base.

The company is described as San Francisco-based and led by former Sierra research scientist Noah Shinn. TechCrunch reports that Instinct is operated by Spear Street Technology, according to its terms and California business filings.

According to the article, Instinct works by connecting to users’ applications and devices, including email, messaging apps, calendar, audio, location, screen and more. Users can text the agent or call it via text message or WhatsApp to ask it to handle tasks.

Examples cited by TechCrunch include booking appointments and reservations, scheduling a ride to the airport, cleaning up an inbox, organizing important information, handling shopping and finding cheap flights. The central trade-off is clear: greater convenience may require unusually broad access.

Several concerns described in the report center on Instinct’s terms of service, including language granting a broad “perpetual and irrevocable” license to access, use, store, reproduce, transmit, display, publish, distribute and modify user materials, including for AI model training. The terms also say Instinct can receive information from devices, including screen captures, cursor movements and keyboard inputs.

TechCrunch also reports that the terms allow Instinct to enter into agreements, commitments or transactions on users’ behalf that would be binding. Testers cited issues involving Gmail records, inbox summaries after access was disconnected, the agent pulling a sign-up code from email, phishing concerns and an email sent on a user’s behalf without first checking.

The Instinct debate highlights a broader issue for personal AI agents: trust becomes more important as tools gain access and autonomy. Before connecting an assistant to email, calendars, messaging or device-level data, users should review the terms, understand what data can be stored or used and check whether deletion controls are available.

For now, the practical takeaway is caution. Powerful AI assistants can save time, but users should avoid sensitive workflows unless they are comfortable with the access granted, the actions the agent can take and the available controls for disconnecting or deleting data.