OpenAI safety researcher David Robinson resigns, warns AI risks need stronger safeguards

David Robinson, a former OpenAI safety employee, has resigned, warning that the company's rapid AI development risks serious failures and that the "time for trial and error is over." Robinson, who spent 3-1/2 years at OpenAI drafting its preparedness framework and overseeing safety reports for 12 model launches, said the company's sprint between releases undermines necessary caution. He argued that advanced AI needs safeguards comparable to nuclear power and aviation, not iterative deployment. His resignation, published in The Atlantic, adds to industry debate over whether AI capabilities are outpacing control and understanding. OpenAI defended its approach, stating it can pause training or hold back models when needed to ensure safety.