1Password says it can fix login security for AI browser agents

The 1Password browser extension autofills your passwords when you browse, and now the company has built a similar tool for AI bots that browse the web for you, but for a completely different reason.

AI tools and browsers built on Claude, Gemini, and ChatGPT are increasingly using AI agents to browse the web, book tickets, and create Spotify playlists for you, and unlike the risk of forgetting a unique password, an AI bot runs the risk of remembering it and causing a hack later. 1Password’s fix for this potential risk is its new Secure Agent Autofill feature that “enters credentials directly into the browser if, and only if, a human approves access.”

Using the tool, when the browser’s AI agent determines that it needs login credentials, the agent “informs 1Password that the credentials are being requested,” Says 1Password. “At this point, 1Password identifies the appropriate credentials, and requests consent from the user via a human-in-the-loop workflow.” To approve a request, a human authenticates the request using something like Touch ID on their Mac, and 1Password uses “End-to-end encrypted channel” Between its browser extension managed by the AI ​​agent and the consent device for entering credentials. According to 1Password, the AI ​​agent and LLM never see the actual credentials as a result.

First, Secure Agentic Autofill is available starting today for early access via Browserbase, which builds a browser and tools specifically designed for AI agents.