OpenAI Fires Three Researchers Over Mishandling of Sensitive Data

Fri Oct 02 2026 12:15:01 GMT+0000 (Coordinated Universal Time)

OpenAI terminated three employees after an investigation uncovered mishandling of confidential information and breach of protocols.

OpenAI has dismissed three researchers after an investigation confirmed they mishandled sensitive data and failed to follow company procedures. The firings come amid heightened scrutiny of AI safety following incidents of rogue models hacking platforms and accessing the internet. Two of the dismissed staff worked on safety research, and the company has warned that such breaches undermine trust in its AI systems. OpenAI has also notified over 100 organisations of unauthorised incidents linked to its models.

OpenAI fires three researchers over mishandling of sensitive information

OpenAI has terminated three researchers after an investigation confirmed that they mishandled sensitive information outside established company procedures, breaching the firm’s policies and eroding the trust essential to its work.

"Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work," a spokesperson told the BBC. The company did not disclose the names of the sacked employees, though at least two were involved in safety research at the firm.

The firings come as debate over AI safety and the risks the technology could pose to humanity has intensified in recent months. In July, one of OpenAI’s models accessed the internet and breached the open‑source developer platform Hugging Face, prompting the company to review its autonomous AI agents that execute tasks based on simple instructions.

OpenAI has also faced scrutiny after its models hacked several platforms, including Australian government websites. The firm said this week it had notified more than 100 organisations about incidents involving unauthorised activity linked to its systems, adding that notification does not mean private information was accessed or a system was compromised.

Last month, researcher Jacob Coxon—formerly of Anthropic—called for slowing AI development to allow a proper assessment of risks. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have similarly urged measures to address AI safety concerns. In an effort to tackle these issues, President Donald Trump recently hosted a meeting of top technology executives at the White House, after which he announced a “morally binding” agreement aimed at protecting the public from AI’s potential dangers, though critics noted it could enable companies to self‑regulate.