OpenAI has dismissed three safety researchers who reportedly shared sensitive company information with outside artificial intelligence safety groups, according to a Wall Street Journal report.

The latest round of firings adds fresh fuel to an ongoing debate about how the company handles internal dissent and whether researchers who raise concerns about safety practices have a viable path to do so within the organization.

OpenAI Reportedly Fired Three Safety Researchers

According to the Wall Street Journal, which first reported the departures, OpenAI terminated three employees from its safety research division after an internal investigation concluded they had passed sensitive information to external parties. Engadget and Gizmodo both confirmed the firings through separate reporting, with both outlets noting that the affected employees were not named publicly by the company.

OpenAI said in a statement that it takes the safeguarding of its research and proprietary information extremely seriously, and that unauthorized sharing of sensitive materials constitutes a violation of its policies and employment agreements regardless of the recipient. The company framed the firings as consistent enforcement of its confidentiality standards rather than as a suppression of safety concerns specifically.

These 3 Employees Allegedly Shared Info to Third-Parties

The three researchers, according to the WSJ report, did not leak the information to media outlets or to external competitors. Instead, the information was shared with organizations that focus specifically on AI safety research and advocacy, groups whose work the researchers apparently believed would benefit from having a clearer picture of what was happening inside one of the world’s most influential AI laboratories.

That distinction matters to critics of the firings, who have drawn a sharp contrast between leaking to the press or to rivals and seeking outside safety oversight, arguing the two are fundamentally different actions with very different motivations.

Confidential Data Shared to AI Safety Groups

The specific content of what was shared has not been disclosed publicly by OpenAI, and the Wall Street Journal’s report did not reveal what materials were involved beyond describing them as sensitive internal information.

What is known is that the recipients were not random individuals or commercial competitors but rather established safety organizations operating in the AI governance space.

That framing has led some observers, including researchers outside the company, to argue that the firings could have a chilling effect on the broader field of AI safety research, since it potentially closes off an informal channel through which outside experts have historically been able to gather real-world information about how frontier AI labs operate internally.

OpenAI’s AI Safety Controversies

These firings do not happen in isolation. OpenAI has faced a sustained series of safety-related controversies and high-profile departures over the past two years. In 2024, the company fired Leopold Aschenbrenner, a member of its Superalignment team, after he shared a safety-related document with three external researchers for feedback, a practice he said was common at the company at the time.

That same year, Jan Leike, who co-led the Superalignment team alongside Ilya Sutskever, resigned and publicly stated that safety culture had taken a backseat to shiny products, and that his team had struggled repeatedly to secure the compute resources needed to complete critical safety work.

The Superalignment team itself, which OpenAI had created in mid-2023 with the stated goal of solving the alignment problem for superintelligent AI within four years, was dissolved in 2024. In early 2026, a wave of departures from both OpenAI and Anthropic drew renewed attention to the question of whether safety researchers inside frontier labs have adequate protections and pathways for raising concerns.

OpenAI responded to that wave by stating that it had some of the strongest safety safeguards in the industry, though critics pointed to the pattern of departures as evidence that those assurances were difficult to verify from the outside.

Sam Altman has also now publicly committed to what he has described as pacing the frontier of AI development, agreeing with Anthropic CEO Dario Amodei’s call for the industry to slow down. That commitment, coming as it does at the same moment the company is firing researchers who reportedly sought outside safety input, has struck some observers as contradictory.

Frequently Asked Questions

What Kind of Data Were Shared?

OpenAI has not publicly disclosed what type of information the three fired employees allegedly shared with external safety groups. The company characterized the materials as sensitive internal information without specifying whether they involved model capabilities, safety testing results, internal communications, research findings, or something else entirely.

The lack of specificity has made it difficult to independently assess whether the shared information posed a genuine security risk or whether it was primarily the act of sharing that violated company policy rather than the nature of the content itself. This mirrors the pattern seen in the 2024 Aschenbrenner case, where the fired researcher maintained that the document he shared had been redacted of genuinely confidential details and that the information itself was not the real issue.

How Will This Affect OpenAI?

The practical consequences of the firings depend heavily on how the AI safety community and OpenAI’s own remaining safety researchers respond to them. If the firings discourage current employees from raising concerns informally with outside experts, it could limit the informal oversight that has historically supplemented official regulatory frameworks, which remain incomplete in most jurisdictions.

OpenAI’s IPO ambitions, which have been in motion throughout 2026, may also be affected, since potential investors and institutional partners have begun scrutinizing the governance and safety culture of frontier AI companies more carefully as the industry matures and regulatory attention increases. OpenAI has consistently maintained that its internal safety processes are robust, but the recurring pattern of departures and firings tied to safety disagreements has made that argument harder to sustain publicly.

Were the Employees Named?

None of the three fired researchers have been publicly identified in the reporting from the Wall Street Journal, Engadget, or Gizmodo as of this writing. OpenAI confirmed the firings but did not identify the employees or provide details about what information they had shared or with which specific external organizations.

This stands in contrast to some earlier cases, such as the Aschenbrenner and Izmailov firings in 2024, where the identities of the affected researchers eventually became publicly known through their own statements and reporting from The Information. Whether the current three employees will come forward to share their accounts remains unclear at this stage.