Open Letter urges stronger AI safety protocols and protections for whistleblowers

A group of current and former employees from leading AI companies, including OpenAI and Google DeepMind, issued an open letter highlighting the industry’s lack of safety oversight and calling for stronger protections for whistleblowers. The letter, signed by 11 current and former OpenAI workers and two current or former Google DeepMind employees, emphasises the need for a “right to warn about artificial intelligence.”

The letter criticises AI companies for holding substantial non-public information about their systems’ capabilities, protective measures, and risks, while having weak obligations to share this information with governments and none with civil society. The signatories argue that companies cannot be relied upon to voluntarily share this crucial information.

OpenAI responded by defending its safety practices, highlighting its internal reporting mechanisms and commitment to only release new technology with appropriate safeguards. Google did not provide immediate comments.

The rapid advancements in AI have intensified concerns about its potential harms, leaving regulators struggling to keep pace. Despite public commitments to safety, researchers and employees have pointed out a lack of oversight as AI tools potentially exacerbate existing social harms or create new ones.

The letter, first reported by the New York Times, calls for increased transparency and accountability, including protections for workers raising safety concerns. It suggests prohibiting companies from enforcing non-disparagement agreements that block employees from discussing AI risks and establishing a mechanism for anonymous reporting to board members.

The letter also addresses aggressive tactics by companies like OpenAI to prevent employees from speaking out, referencing a report by Vox about restrictive non-disclosure agreements. OpenAI’s CEO, Sam Altman, has since apologised and promised changes.

This open letter follows the resignations of two top OpenAI employees, co-founder Ilya Sutskever and safety researcher Jan Leike, who criticised the company for prioritising product development over safety.