David Robinson, who said he spent three and a half years at OpenAI, resigned and wrote in The Atlantic that the company's iterative deployment culture is not careful enough. OpenAI said it pauses training or holds models when needed.
A former OpenAI safety employee who recently resigned criticized the company's approach to AI safety in an Atlantic essay published Saturday, Reuters reported. David Robinson wrote that AI companies, including OpenAI, were not being nearly careful enough and should put more weight on safety expertise before building more capable systems.
Robinson said he spent three and a half years at OpenAI, helped draft its preparedness framework and oversaw safety reports for 12 frontier-model launches. He argued that advanced systems need safeguards closer to those used in nuclear power and aviation, and wrote that "the time for trial and error is over."
He said OpenAI relies heavily on iterative deployment, releasing systems and strengthening safeguards when problems emerge. "As the company sprints from one launch to the next, it is failing to achieve the level of care that I believe is needed," he wrote, according to Reuters.
An OpenAI spokesperson said the company is making sure models do not become more capable than it can safely manage and secure, and that it pauses training or holds back models when it needs to slow down.
Robinson also warned that capabilities were advancing faster than researchers' understanding of alignment. Reuters did not report a departure date beyond the recent resignation described in the essay.
People Sentiments Mixed
- Robinson wrote that the time for trial and error is over and that OpenAI is not achieving the level of care he believes is needed.
- An OpenAI spokesperson said the company pauses training or holds back models when it needs to slow down.