Chrome Fortifies Agentic AI Security

Google is implementing new security measures for Chrome's upcoming agentic AI features, powered by Gemini. These features will allow the browser to perform multi-step tasks, like filling out forms and placing orders, on behalf of the user. The new security architecture is designed to protect users from potential risks associated with AI-driven browsing.

Gemini Nano, a smaller AI model, will work directly on devices to block scams, reduce unwanted notifications, and simplify password management. Chrome will also use Gemini to detect scams in real time, identifying fake virus warnings and phishing attempts. These features will be available through Google Workspace with enterprise-grade data protections.

These updates arrive as other browsers are also adding AI features. Google's approach focuses on avoiding early mistakes and ensuring the security of AI agents. The integration of Gemini aims to streamline workflows and reduce friction in digital tasks, while also addressing concerns around data residency and potential misuse.