Story perspectives

1Password launches AI-agent logins without exposing credentials

By Drooid · Oct 22, 2026, 10:06 PM UTC · How we work

Summary

1Password Enables AI Logins

1Password launched 1Password for Agents, allowing AI agents to log in without seeing credentials.
Initially it supports Meta’s Muse, SpaceX AI’s Grok Bot, Claude Duet and OpenAI’s ChatGPT Work.
After a user approves, 1Password injects the password or MFA code for that session.
The feature adds no extra fee for Individual and Family accounts but is unavailable in the EU and Canada.

Summary

Researcher Shows AI Hack

U.S. banks detected AI-agent activity inside customer accounts, prompting a Chase spokesperson to call it “horrifying.”
Security firms are scrambling to develop safeguards against credential-exposing AI agents.
1Password’s annual recurring revenue rose 90% year-over-year, exceeding $500 million.
Monthly active human users of 1Password’s developer tools grew 200%, while non-human identity users surged 500%.
A researcher documented an AI-agent hack that breached over 100 websites, showing abuse potential.

History

Historical Context

The 2017 Equifax breach, discovered July 29 and disclosed September 7, exposed 147.9 million U.S. records and led to a $575 million FTC settlement and a $700 million total penalty. In February 2020 the U.S. indicted Chinese PLA hackers for the attack. Today, 1Password’s AI-agent rollout and a 500 % surge in non-human credential use in 2024 revive those security concerns, prompting stricter authentication rules.

1 of 3

1Password Enables AI Logins

1Password launched 1Password for Agents, allowing AI agents to log in without seeing credentials.
Initially it supports Meta’s Muse, SpaceX AI’s Grok Bot, Claude Duet and OpenAI’s ChatGPT Work.
After a user approves, 1Password injects the password or MFA code for that session.
The feature adds no extra fee for Individual and Family accounts but is unavailable in the EU and Canada.

1 / 3