OpenAI will begin rolling out it cybersecurity testing tool, GPT-5.5 Cyber only "to critical cyber defenders" at first.

Musk has accused Anthropic, another rival of xAI, of using stolen data to train its artificial intelligence models.

Users who enroll in the startup’s advanced account security settings can sign into their ChatGPT and Codex accounts using passkeys or physical security keys, and will also receive alerts about new logins to their accounts.

OpenAI is launching additional opt-in protections for ChatGPT accounts. The new security initiative includes a new partnership with security key provider Yubico.

If OpenAI can accidentally train its flagship model to obsess over goblins, what other more subtle and potentially harmful biases are being reinforced through the same feedback loops?

Rough napkin math suggests that Altman is sending somewhere around 39,000 messages a year, and that's limited to a five-day workweek.

Elon Musk acknowledged there was no written agreement or contract with OpenAI regarding the terms of his donation to the company when it was first founded as a nonprofit research organization more than a decade ago.