Former and current employees from OpenAI and Google DeepMind have signed an open letter addressing the lack of transparency and safety oversight in the AI industry. The letter highlights the need for AI companies to share information about risks, protect whistleblowers and foster a culture of open criticism.
A group of current and former employees from OpenAI and Google DeepMind issued an open letter on June 4, 2024, addressing the need for greater transparency and safety oversight within the artificial intelligence (AI) industry. Eleven individuals from OpenAI and two from Google DeepMind signed the letter, which highlights the risks associated with AI technologies and calls for protections for whistleblowers.
The letter asserts that AI companies often possess significant non-public information about their systems’ capabilities and limitations, yet they have weak obligations to share this information with governments and none with civil society. The signatories are urging for a culture that welcomes open criticism and protects employees who voice concerns about potential risks, which they believe companies currently have strong financial incentives to avoid.
OpenAI responded by affirming its commitment to safety and transparency, noting that it has mechanisms such as an integrity hotline and a Safety and Security Committee to address issues. Despite these assurances, former OpenAI engineer Daniel Ziegler expressed skepticism about the company’s pledge to transparency, stressing the importance of a culture that allows employees to raise concerns.
The letter also asks AI companies to commit to four principles surrounding transparency and accountability, including allowing employees to discuss risk-related issues without fear of retaliation. This comes in the context of a rapidly advancing AI field where regulators are struggling to keep pace with technological developments.
Additionally, the open letter follows recent reports of restrictive non-disparagement agreements at OpenAI, actions that CEO Sam Altman has since pledged to amend. The letter emphasizes that, in the absence of effective government oversight, current and former employees play a critical role in holding AI companies accountable to the public.
