OpenAI Adds Advanced Account Security

2026-05-04 · openai

OpenAI introduced Advanced Account Security, a new opt-in setting aimed at people who need stronger protection against takeover and phishing. The feature extends to Codex as well, making security a first-class concern across OpenAI's developer workflow. The release is also a practical reminder that AI accounts can hold sensitive personal and professional context, so recovery and session controls matter more than ever.

Key Features or Updates

Advanced Account Security requires passkeys or physical security keys, disables password-based login, and removes email/SMS recovery. It also shortens sessions, adds login alerts, and can automatically exclude protected conversations from training.

Impact on Developers

For teams using ChatGPT and Codex on sensitive work, this reduces one of the biggest operational risks: account compromise. It also makes phishing-resistant auth easier to standardize for high-trust workflows.

How to use it

Enroll from the Security section of ChatGPT on web, then switch to passkeys or a FIDO security key. If your workflow depends on Codex or sensitive prompts, this is worth enabling early rather than after an incident.

Read Original Post →