Everyone is buzzing aboutInstinct, an AI personal assistant still in private access, not only for its incredible capabilities, but also for its potential privacy concerns. The agent, a veritable taskmaster, has beenpraisedas feeling “like magic” andbeingone of the “most exciting launches” since OpenClaw. However, some testers have also raised concerns about the AI agent’s security model and its worrisometerms of service.
To be fair, Instinct is still in private testing for now, so these concerns haven’t yet scaled to the wider public at this time.
Created by a small team led by former Sierra research scientistNoah Shinn, San Francisco-based Instinct is operated bySpear Street Technology,per itstermsandCaliforniabusinessfilings. It’s currently operating in stealth, perPitchBook.
The AI agent itself works by connecting to your applications and devices, including your email, messaging apps, calendar, and your device’s audio, location, screen, and more. You can text the agent or call it via text message or WhatsApp, asking it to perform various tasks for you, like booking your appointments and reservations, scheduling a ride to the airport, cleaning up your inbox, organizing important information, handling your shopping,findingyou cheap flights, and much more.
Though Instinct is described by testers as outperforming their expectations, some have also raised concerns about the company’s approach to customer privacy and security. This raises a timely question: are the trade-offs of giving AI this level of access and autonomy actually worth it?
For instance, several peoplearecirculatingscreenshotsfrom the company’sTerms of Service, whichgrantInstinct a broad “perpetual and irrevocable” license to “access, use, host, cache, store, reproduce, transmit, display, publish, distribute, and modify” any of the user’s materials, including for training its AI models. The terms also detail how Instinct can receive information from users’ devices, including screen captures, cursor movements, and keyboard inputs.
The terms also allow Instinct to enter into “agreements, commitments, or transactions” on users’ behalf, which would be binding.
One early adopter, Peter Yang,pointed out thatInstinct would not delete his Gmail records when asked. (The team laterfixedthe problem by adding a tool for deleting external data in its settings, he said.)
Another person, Claire Vo, found that Instinct was still summarizing their inboxafter disconnecting its access. When she asked Instinct what happened, the bot confirmed that the emails were stored in plain text for later searches.
Many others questioned the security model, too. One testerwas a bit worriedwhen they found that Instinct was able to pull a sign-up code from their email inbox to complete a particular task — in their case, booking a table at a restaurant via Resy. And Hello Patient co-founder Alex Cohen wrote that once he found outhow easily Instinct could be phished, he deleted his account:
“We’re trading privacy and control for hyper-personalized AI tools (AI notetakers, personalized AI agents, etc), often without fully understanding the trade,” she remarked on X, summarizing the dilemma posed personal AI agents. “The more powerful these agents become, the more trust matters. Every successful action earns a little more trust. One unauthorized action can reset that trust to zero,” she said.
Michael Mignano, the founder of Anchor, which was acquired by Spotify, and now a GP at Union Square Ventures,noted thatproducts like Instinct are going to “change modern security norms for consumers,” adding that “people will increasingly hand over passwords to 3p [third-party] apps, unaware of how or what they are storing for them.”