Three Fired OpenAI Safety Researchers Warn Dismissals Could Chill AI Safety Work

Three recently dismissed OpenAI safety researchers have published a letter warning that their firings could chill the internal debate and outside collaboration needed to manage increasingly autonomous AI systems. Tomek Korbak, Jasmine Wang, and Mikita Balesni say OpenAI cannot make AI safe on its own and urge the company to keep deep third-party safety access, preserve model monitorability, and clarify how employees may work with outside safety organizations.