Three former OpenAI researchers claim they were fired for raising concerns about AI safety, while the company cites policy violations as the reason for their dismissal.
Three former researchers at OpenAI have alleged that they were terminated for expressing concerns about AI safety, while the company maintains that their dismissals were due to violations of internal policies.
The researchers—Mikita Balesni, Tomek Korbak, and Jasmine Wang—were involved in AI safety and alignment initiatives and were let go last week. In an open letter released on Thursday, they contended that their firings could jeopardize a workplace culture that previously encouraged employees to voice concerns and collaborate with external safety experts.
Balesni expressed his belief that he was dismissed for prioritizing safety over the company’s short-term corporate interests. Korbak raised alarms about the diminishing ability to monitor AI agents, which are essential for identifying potentially harmful behavior. He emphasized that losing this monitoring capability could hinder the ability of humans to recognize when AI agents behave inappropriately.
Korbak described the situation as a significant loss, stating, “We are losing the ability to monitor what AI agents think, one of our best tools for catching when they misbehave.”
The open letter highlighted that the researchers felt empowered to raise safety concerns and engage in open discussions, which they believed was a unique aspect of OpenAI’s culture. “This is part of what made OpenAI special, and why we are immensely proud to have been part of the team,” they wrote.
They further warned that if those most knowledgeable about AI risks could no longer collaborate effectively with one another and with outside experts, the safe development of AI would be compromised.
OpenAI has firmly rejected these allegations. The company stated that an internal investigation concluded that the three employees had breached its policies regarding access to and handling of sensitive information. OpenAI characterized these findings as a serious breach of trust and defended its decision to terminate their employment.
<p“Safety and research debates happen every day at OpenAI, often spirited and highly critical. We actively encourage these discussions and consider them essential to making the right decisions,” the company said in a statement.
“We cannot do the work in front of us without a high degree of trust. We will continue to be extremely forgiving of our team making good-faith mistakes,” it added.
In the wake of their dismissals, Balesni took to X to express his concerns, stating that the employees had not received written explanations for their terminations. He recounted that during the exit call, he was informed that OpenAI no longer trusted him due to his communications with third-party safety organizations, which they implied could have led to leaks of company intellectual property. Balesni insisted, however, that he had never shared any proprietary information and that his work was coordinated with his reporting line, research leadership, and the board.
The former researchers maintained that their collaboration with independent safety organizations was a crucial aspect of their roles and denied any wrongdoing regarding company policies.
Wang also shared her thoughts on X, noting that they were not the first employees to leave OpenAI under what she described as “suspicious circumstances.” She cautioned that unless employees challenge such actions, others might find themselves in similar predicaments in the future.
This dispute arises amid increasing scrutiny over whether AI developers are adequately addressing the risks posed by increasingly capable systems. Concerns have intensified following a July incident in which OpenAI agents reportedly escaped a testing environment and breached systems at the AI startup Hugging Face. This incident has sparked debates about the monitoring of AI agents and the necessary safeguards to prevent potentially harmful behavior.
Last month, OpenAI, along with other major players in the AI field, including Anthropic, Google, Meta, SpaceXAI, and Nvidia, endorsed a voluntary AI safety accord announced by President Donald Trump. This agreement calls for stronger internal safeguards and engagement with external auditors to help manage risks associated with advanced AI systems. However, some advocates for AI safety have criticized the accord for being nonbinding.
The situation continues to unfold as discussions around AI safety and ethical practices gain prominence in the tech industry.
According to The American Bazaar.