browser-using AI is inevitable. But it brings serious safety and security challenges that we need to solve properly. So we’re starting with an experiment. We’ve built Claude for Chrome, where you can chat with Claude in a sidebar, and have it take actions on your behalf within your browser window. What Claude can do in Chrome: * Manage your calendar and email * Handle routine tasks like data entry or repetitive forms * Test websites by navigating user flows We’re piloting this with a select group of Max users to gather real-world data on use cases, limitations, and safety issues, before expanding access. Browser AI tools are already emerging across the ecosystem, many of which are built on Claude’s models, making safety improvements in this area urgent. Important: Browser use is experimental technology with real risks. We’ve built permission controls and safeguards, but start with trusted sites and supervise Claude carefully.Learn more. Join the research preview waitlist Once you have access, you’ll get installation instructions and safety guidelines. - The Anthropic Team   This email was sent to pfeifferj@archlinux.ch. To opt out of future emails, clickunsubscribe . Anthropic PBC, 548 Market St, PMB 90375, San Francisco, CA 94104 [Research preview] Help us develop safe browser AIAnthropic Team undefinedpfeifferj@archlinux.ch pfeifferj@archlinux.ch“T‚ÚP