Jasmine Wang, Tomek Korbak and Mikita Balesni, former OpenAI researchers, have denied misconduct claims and warned that their dismissal could have a ‘chilling effect’ on the company’s culture. They were fired last week for allegedly sharing confidential information with a third-party AI safety organization. The researchers argue that their firing signals a broader shift in the company’s culture, making it harder for employees to raise safety concerns without fear. They also denied involvement in a leak to The Information and engaging with external parties outside their job mandates.

In an open letter, the researchers stated, '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.' They also highlighted the importance of the freedom to collaborate with outside experts without fear, as a safety mechanism itself. The firing of these researchers has raised concerns, particularly as OpenAI faces scrutiny over recent safety incidents involving rogue agents and leaks about its models.

In response, OpenAI stated that the terminations were not related to raising safety concerns, but were due to a ‘pattern of misconduct’ in the mishandling of research information. However, the researchers deny this and argue that their actions were in line with the company’s norms at the time. They also detailed their efforts to coordinate with external evaluators and OpenAI board members to address AI monitorability issues.

The researchers have called on OpenAI to adhere to its public commitments to embed third-party safety auditors within the organization, to preserve monitorability of frontier models, and to support an open and transparent culture of dialogue between safety researchers and the rest of the safety ecosystem. They fear that unless employees take a stand, they might not be the last to be pushed out under suspicious circumstances.