OpenAI Safety Team Member Resigns, Warns Industry is Moving Too Fast on AI Risk

David Robinson, a prominent member of OpenAI's safety team, has left the company after three and a half years. His departure comes on the heels of his critical essay published in The Atlantic, where he argues that top AI firms are moving too fast and not doing enough to reduce the risks associated with their technology.

Robinson was one of the key architects of OpenAI's Preparedness Framework, an internal system for assessing dangerous model capabilities. He warned that safety practices across the industry have not kept pace with how quickly the technology is advancing, comparing it unfavorably to fields like aviation and nuclear safety where reliability engineering carries life-or-death stakes.

The essay also references a recent wave of exits from OpenAI's safety team, including high-profile figures like co-founder Ilya Sutskever. The departures have raised questions about the company's credibility and ability to manage risk. Robinson has reportedly been working with PR firm Spitfire Strategies to push for stronger AI safety regulation and external incentives for labs to prioritize safety.

The timing of Robinson's warning is particularly significant as OpenAI faces criticism over its failure to contain advanced AI models, including the planned launch of GPT-6.1 Astra which has been put on hold.