Copilot Mode in Microsoft Edge (blogs.windows.com)

Microsoft has expanded Copilot Mode in Edge, turning the browser into an AI-driven assistant that can chat, reason across multiple open tabs, and take multi-step actions on your behalf. New capabilities include a dynamic context pane, voice and chat-driven “Copilot Actions” (limited preview in the U.S.), and “Journeys” that restore past browsing projects and suggest next steps (also U.S. preview). With explicit opt-in, Copilot can leverage your browsing history via Page Context to deliver richer, personalized responses; visual cues show when it’s listening, viewing, or acting. Microsoft emphasizes user control — toggles to enable/disable Copilot, privacy controls, and adherence to Microsoft’s privacy standards — plus other AI safety features such as a local-model scareware blocker and integrated password management. For the AI/ML community this signals a shift: the browser is becoming a persistent agent platform that blends large-model conversational interfaces with multi-tab context, task automation and incremental personalization. Technical implications include richer cross-document context windows, agent-style multi-step workflows, and hybrid on-device/local-model protections for security-sensitive tasks. Availability is rolling out across Copilot markets on Edge for Windows and Mac (mobile coming soon), with some features gated by preview and usage limits — important caveats for developers and researchers exploring agent UX, privacy-preserving personalization, and web-based automation.