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

Instinct is an AI personal assistant still in private testing that connects to users' applications and devices, including email, messaging apps, calendars, and device audio and location, to perform tasks like booking appointments and organizing inboxes. The tool has been praised by testers for powerful capabilities, but concerns have emerged about its privacy and security practices, including broad terms of service granting the company rights to access, store, and use user materials for training, as well as incidents where the AI accessed sensitive information like sign-up codes and sent emails without explicit user permission. Early adopters have reported security vulnerabilities, including the AI being susceptible to phishing attacks and retaining data even after access was supposedly disconnected. The concerns raise fundamental questions about whether the convenience of granting AI this level of access to personal information and decision-making authority is worth the privacy and security trade-offs involved.