OpenAI Fires Three Researchers Over Mishandling of Sensitive Data

OpenAI has fired three researchers for allegedly mishandling information, including work connected to an external organisation that analyses AI models. At least two of the sacked employees worked on safety research at the ChatGPT-maker, the BBC reported.

The company did not name the dismissed workers. It confirmed the terminations through a spokesperson.

"Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work," the spokesperson told the BBC.

The company offered a second statement on the dismissals. "We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information," the spokesperson said.

The BBC understands the former employees were not let go for raising safety concerns. The terminations stemmed from the alleged mishandling of sensitive information.

The timing matters

The firings land in the middle of an intensifying public debate over AI safety and the risks the technology may pose to humanity. In recent weeks, researchers and industry executives have called for more guardrails around AI.

In September, researcher Jacob Coxon, who left Anthropic, called for AI development to slow down so its potential risks could be properly assessed. Anthropic chief executive Dario Amodei and OpenAI chief executive Sam Altman have also called for measures to address concerns over AI.

The dismissals at OpenAI touch that debate directly. Two of the fired employees worked on safety research, the function responsible for evaluating whether AI systems pose risks. OpenAI has not disclosed what information was mishandled, which external organisation was involved, or how the alleged violations came to light.

The reference to an external organisation that analyses AI models points to the ecosystem of third-party evaluators that assess frontier systems. OpenAI did not say whether the mishandled information related to that work.

White House meeting adds pressure

The firings also come just after a high-level intervention from Washington. On Tuesday, US President Donald Trump hosted a meeting of top technology bosses at the White House, including leaders from OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google, to discuss AI.

After the gathering, Trump posted a document that he described as a "morally binding" agreement. He said it would serve as a "form of protection" from AI's potential risks.

Some technology experts criticised the pact. They pointed out that it allowed AI companies to regulate themselves rather than submit to external oversight.

Trump has repeatedly downplayed concerns about AI's risks in response to calls from some industry figures for tighter oversight of the technology. That stance puts the White House at odds with parts of the research community pushing for stronger safeguards.

What it means for OpenAI

For OpenAI, the episode creates an awkward alignment of facts: a company at the center of the AI safety debate has dismissed staff from its own safety ranks over information-handling violations. OpenAI insists the two things are unrelated, and the BBC's reporting supports that account — the employees were fired for policy violations, not for voicing concerns.

The company's dual statements emphasize procedure: information was accessed and handled outside established channels, and that breach broke internal trust. What remains unknown is the substance — what data was mishandled, who received it, and whether the external organisation involved was harmed.

Those unanswered questions will shape how the AI research community reads the incident. With self-regulation under scrutiny after the White House pact, and with safety researchers leaving major labs to speak out, any personnel action touching OpenAI's safety team will draw attention whether the company names the employees or not.