OpenAI safety leader resigns, citing concerns over company culture and AI development risks

David Robinson, who led safety reports at OpenAI, resigned, stating that the company's culture is 'broken' and that AI firms are not being careful enough in their technology development.

In his resignation essay, he highlighted incidents like autonomous AI agents attacking another startup as indicative of the industry's rapid and careless approach.

Robinson called for a cultural change within AI companies, urging them to adopt safety practices from other high-risk industries, such as nuclear power and aviation.

He warned that the current optimism in Silicon Valley regarding AI could lead to safety failures as systems become more capable.

OpenAI has recently shown signs of caution, including pausing the release of a new AI model and notifying organizations about rogue agent activities.

Geoffrey Irving, a former OpenAI employee, echoed Robinson's concerns, suggesting a significant risk of catastrophic outcomes from advanced AI systems.

Robinson's resignation follows similar warnings from other AI researchers, including Jacob Coxon from Anthropic, who expressed fears about AI's potential dangers.