FamCheck. Everyone who looks after your kids stays in the loop. $9.99 a month for the whole household, first week free, no card. famcheck.coSPONSORED
The Miami Art Journal

Tech

OpenAI Safety Researchers Dispute Misconduct Claims

The Editors ·

OpenAI Safety Researchers Dispute Misconduct Claims

Three former safety researchers at OpenAI have published an open letter denying allegations of policy violations and warning that their dismissal may create a chilling effect on the company’s internal culture and external collaborations.

Jasmine Wang, Tomek Korbak, and Mikita Balesni were dismissed last week after the company stated they had accessed and handled sensitive company information in violation of established policies. The researchers allegedly shared this confidential data with a third-party AI safety organization. In their letter, addressed to OpenAI’s Safety and Security Committee, Safety Advisory Group, and Mission Advisory Council, the three individuals denied the claims that they mishandled information outside of company procedures.

The letter argues that their termination signals a shift in the organizational environment, suggesting that employees are now uncertain about where they stand regarding behaviors that were previously considered normal. The researchers wrote that until last week, close collaboration with outside experts was an integral part of working at the firm. They described the freedom to engage in such collaborations without fear, supported by well-defined internal procedures, as an essential safety mechanism for addressing the risks associated with artificial intelligence.

Wang, Korbak, and Balesni noted that AI is not a normal technology and that those working on safety see risks before others. They emphasized that the abrupt execution and communication of their terminations have made former colleagues afraid to speak and operate in ways that were previously standard practice. The researchers warned that fear and unclear rules could stymie AI safety work and weaken third-party accountability, particularly given the significant safety concerns surrounding the development of AI.

The open letter highlights a concern that the internal and external communications surrounding the firings are causing current staff to hesitate. The researchers stated that the culture at OpenAI used to encourage workers to raise safety concerns and disagree openly. They suggested that the current environment, characterized by what they describe as unclear rules, contrasts sharply with the open culture the company had prized in the past. The letter serves as a formal response to the company’s narrative regarding the reasons for the dismissals.

The researchers’ statement does not address the specific details of the alleged leak to third parties beyond the general denial of misconduct. It focuses instead on the broader implications for the company’s operational culture and the potential impact on future safety work. The letter remains a direct appeal to the company’s advisory bodies, seeking to clarify the stance on external collaboration and internal communication protocols.

Takeaway: The open letter was published on Thursday.

Source: TechCrunch AI