David Robinson, an OpenAI employee responsible for artificial intelligence safety, has resigned, warning that top AI companies are still inadequate in risk reduction. Robinson was in charge of transparency work on OpenAI's safety team. In an article for The Atlantic, he pointed out that the developer of ChatGPT has been relying on a trial-and-error approach for development. He stated that despite the intelligence and hard work of his colleagues, the company has failed to meet the required standards of caution during the rapid iteration of its products. Robinson emphasized that what the field of artificial intelligence truly needs is a rigorous attitude of 'near perfection on the first attempt.'
