A group of current and former employees from prominent AI companies issued an open letter on Tuesday warning about the industry's lack of safety oversight and calling for enhanced whistleblower protections.
Signed by eleven current and former OpenAI employees and two from Google DeepMind, the letter advocates for a "right to warn about artificial intelligence" and criticises the industry's secrecy.
It also highlights the substantial non-public information AI companies possess about their systems' capabilities and risks, noting their weak obligations to share this with governments and none with civil society. It argues that companies cannot be relied upon to voluntarily disclose such information.
In response, OpenAI defended its safety practices, citing avenues like a tipline for reporting issues and stating that new technologies are not released without safeguards. Google did not comment immediately.
The letter calls for protections for workers raising safety concerns, including transparency principles and the elimination of non-disparagement agreements that prevent airing AI risk-related issues.