
David Robinson, a former safety employee who spent three and a half years at OpenAI, resigned and published an essay criticizing the company's fast-paced development culture.
Robinson argued that AI companies are not exercising sufficient caution and need stricter safeguards before launching increasingly capable frontier models.
An OpenAI spokesperson responded by stating that the company pauses training or holds back models when safety management requires slower progression.