OpenAI has dismissed three researchers for mishandling sensitive company information, marking the latest in a series of personnel actions at the artificial intelligence firm related to data security and safety concerns.
The three individuals—identified by news reports as Jasmine Wang, Tomek Korbak and Mikita Balesni—allegedly shared confidential research with an external AI-safety organization, according to reporting from technology outlets. Neither the organization nor the specific information involved has been publicly disclosed.
Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work, an OpenAI spokesperson stated. The company added that it had
parted ways with three individuals for violating our policies on accessing and handling sensitive company information.
At least two of the terminated employees worked in safety research roles at the firm. An investigation by OpenAI uncovered what one news account described as a
pattern of misconductin how personnel with access to confidential data managed company research.
Context within OpenAI's history of personnel actions
This termination follows earlier dismissals at the company. OpenAI fired researchers Leopold Aschenbrenner and Pavel Izmailov in 2024 over alleged information leaks, establishing a pattern of strict enforcement around data handling. The latest action underscores the organization's stated commitment to protecting proprietary research and maintaining internal security protocols.
Why does AI safety matter to OpenAI?
The firings occur amid intensifying public debate about artificial intelligence safety and potential risks the technology poses to society. OpenAI chief executive Sam Altman and Anthropic leader Dario Amodei have both advocated for regulatory measures addressing AI-related concerns. In September, researcher Jacob Coxon, who departed Anthropic, called for AI development to decelerate to allow proper assessment of potential dangers.
On September 22, OpenAI published principles endorsing independent safety assessments with comprehensive access across training, evaluation and deployment phases—a move signaling the company's stated openness to external scrutiny of its safety practices.
What is the broader policy landscape?
On Tuesday, US President Donald Trump convened a meeting of leading technology executives—including representatives from OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google—to address artificial intelligence policy. Following the White House gathering, Trump released a document he characterized as a
morally bindingagreement intended to serve as
a form of protectionfrom AI's potential risks.
Technology experts have raised objections to the agreement, noting that it permits AI companies to oversee their own compliance. Trump has consistently minimized concerns about AI risks in response to calls from certain industry figures for stricter regulatory oversight of the technology.
Key Facts
- Three OpenAI researchers—Jasmine Wang, Tomek Korbak and Mikita Balesni—were terminated for sharing confidential information with an external AI-safety organization
- OpenAI's investigation identified a pattern of misconduct in how employees with access to sensitive data handled company research
- The dismissals represent the latest in a series of personnel actions at OpenAI related to data security, following 2024 terminations of Leopold Aschenbrenner and Pavel Izmailov
- The action occurs as AI safety debates intensify, with industry leaders and policymakers discussing regulatory frameworks and oversight mechanisms
- OpenAI has publicly committed to supporting independent safety assessments with deep access to its training, evaluation and deployment processes
