1Password has introduced a new feature called 1Password for Agents, designed to allow AI assistants to log into users’ personal accounts without ever accessing their actual passwords. This solution addresses growing concerns around security as consumers increasingly entrust AI agents with sensitive tasks. The system works by sending a unique digital key to confirm the AI agent’s identity, prompting the user to approve the login attempt before 1Password fills in the credentials to complete the login securely on behalf of the agent.

This innovative platform supports a range of AI assistants, including Claude Duet from Anthropic, OpenAI’s Dots and ChatGPT Work, Grok Bot, Meta’s Muse, and agents from the startup Town. 1Password is also actively collaborating with other companies to expand compatibility. By enabling users to delegate tasks securely, the company aims to prevent the risk associated with sharing passwords directly, where credentials might be copied and stored beyond the user’s control.

The launch of 1Password for Agents comes amid growing apprehension about AI-driven agentic activity in sensitive areas such as banking. A report from Axios on the same day highlighted the detection of automated agent activity within U.S. bank accounts, which poses challenges because financial institutions often cannot distinguish whether such activity is authorized by the user. These developments underscore the necessity for technologies that strike a balance between convenience and security in the era of AI.

Banks have voiced concern over this trend, warning customers not to share their login credentials casually and describing the current situation as troubling. With 1Password offering a trusted way to manage and grant access to AI agents without exposing credentials, this new platform could become a critical tool for safely integrating AI technology into personal and financial digital lives.