Former OpenAI employee David Robinson urges greater safety focus for AI development

David Robinson, a former safety employee at OpenAI, criticised the company’s current methods regarding artificial intelligence safety practices. Robinson asserted that AI firms, including OpenAI, must prioritise safety research and expertise before advancing more powerful systems.
During his tenure, Robinson spent three and a half years working at OpenAI. He was responsible for overseeing safety reports across twelve launches of frontier models.
OpenAI currently depends significantly on a process it terms “iterative deployment”. Conversely, an OpenAI spokesperson stated that the organisation can slow its pace by pausing training or withholding models as necessary.
David Robinson stated, “The time for trial and error is over”.
David Robinson























































