1Password launches secure sign-in platform for artificial intelligence agents

1Password has launched a new sign-in system that helps personal artificial intelligence (AI) agents safely access people's online accounts without seeing their passwords. The news was reported by Silicon Angle.

The new system aims to fix a security problem that happens when people copy and paste their passwords directly into AI chat windows. Doing this can put private information at risk. Instead, the system uses a special digital key. This lets AI agents from companies such as Meta Platforms Inc., Anthropic PBC, OpenAI Group PBC, and SpaceXSI prove they have permission to access an account without ever seeing the real password.

Before an AI agent can sign in, the account owner must approve the request on their own device. The system then automatically enters the sign-in information. This means the AI agent gets access only for a short time and only for the specific job it needs to do. It does not get permanent control of the account. Large AI companies, including OpenAI and Anthropic, have already started adding this technology to their AI assistants. Nine other companies are also working on adding similar features to make their users safer.

This new system is part of a bigger effort across the technology industry to create rules that let AI agents access accounts only for specific tasks. The goal is to make sure AI agents have only the permissions they need when doing work for people. 1Password serves more than 200,000 businesses, including well-known companies such as Canva Pty Ltd. and Stripe Inc. However, the company has not yet announced exactly when these new connections with AI assistants will become available.

Source: Silicon Angle