1Password Now Lets AI Agents Log In for You—Without Seeing Your Password

1Password has introduced a new feature called 1Password for Agents that allows AI assistants to log into your online accounts on your behalf, but without ever accessing your actual passwords. This solves a growing problem: as AI agents become more common, people often share their passwords with them, which can be risky. With this tool, when an AI agent needs to log in, it sends a request to 1Password. You get an alert on your device and can approve or deny the request. If you approve, 1Password securely logs the agent in, and the agent can complete its task—like booking a reservation or checking your email—without ever seeing your credentials.

The service works with several popular AI agents, including Claude Duet from Anthropic, OpenAI's Dots and ChatGPT Work, Grok Bot, Meta's Muse, and agents from Town. 1Password says it's working with more companies to expand the list. The launch comes as U.S. banks are reportedly detecting AI agent activity in customer accounts, which has raised concerns. Banks say they can't always tell if the activity is authorized by the user. A Chase spokesperson advised customers not to share their usernames and passwords with anyone.

This new tool aims to give people a safe way to delegate tasks to AI agents while keeping control of their passwords. It's like having a trusted assistant who can access your accounts only with your permission each time. The catch is that you still need to approve each login, so it's not fully hands-off. But for those worried about security, that extra step might be worth it.