Fired OpenAI Safety Researchers Contest Dismissals and Warn of Chilling Workplace Effect

Three former OpenAI safety researchers—Jasmine Wang, Mikita Balesni, and Tomek Korbak—have issued an open letter disputing their recent terminations. The trio warned that the high-profile manner of their dismissals risks silencing remaining staff at a critical juncture when artificial intelligence safety is paramount. OpenAI terminated the trio last week, claiming they violated internal policies regarding the handling and accessing of sensitive corporate data and compromised the trust necessary for frontier research.

In their letter, the researchers contested the company's official rationale, stating they acted within established norms and directly in line with the startup's mission. "We have become concerned that internal and external communications around our firing have made our former colleagues afraid to speak and operate in ways that, until last week, were an integral part of working at OpenAI," the letter stated. They also explicitly denied leaking details to the press regarding new, less monitorable system architectures, maintaining their external discussions occurred in coordination with board members and senior executives.

The terminations follow recent high-profile security incidents involving OpenAI agents and have heightened scrutiny from external regulators and lawmakers. The researchers emphasized that their interactions with independent AI safety organizations were previously encouraged to ensure independent oversight. Wang later posted on social media that the dismissals sent an alarming signal to remaining staff, cautioning that employees who raise safety questions or collaborate with outside entities face sudden retaliatory measures.

Moving forward, the researchers urged OpenAI leadership to maintain commitments to third-party safety audits rather than using their dismissals as a pretext to scale back external scrutiny. They also demanded that the company preserve the monitorability of frontier models and protect transparent internal discourse. While OpenAI leadership reportedly expressed agreement with parts of their message, the dispute adds to deepening tensions regarding whether corporate commercialization is eroding the company's original safety-first charter.