David Robinson made the comments in an article titled “I Quit OpenAI Because Its Culture Is Broken”.
David Robinson said the “time for trial and error is over” and argued that AI companies need to put far greater emphasis on safety before developing more powerful systems. A former OpenAI safety employee has resigned from the company, warning that its rapid development of increasingly capable AI systems could raise the risk of serious failures.
Robinson spent 3-1/2 years at OpenAI, where he helped draft the company’s preparedness framework and oversaw safety reports for 12 frontier-model launches.
Robinson made the comments in an article titled “I Quit OpenAI Because Its Culture Is Broken”, published by The Atlantic on Saturday. Robinson said AI companies, including OpenAI, were not being “nearly careful enough” and argued that safety research and expertise should play a much bigger role before more powerful AI systems are developed. He said the company’s focus on moving quickly from one AI model launch to another was making it difficult to maintain the level of caution he believed was necessary. OpenAI refers to this approach as “iterative deployment”, under which systems are released and their safeguards are strengthened as new problems are identified. “The time for trial and error is over,” he wrote, saying advanced AI systems should have safeguards comparable to those used in high-risk fields such as nuclear power and aviation. “We’re making sure our models don’t become more capable than we can safely manage and secure, and we pause training or hold back models when we need to slow down,” an OpenAI spokesperson said in a statement.
“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 argued that the development of advanced AI should no longer rely on a process of repeatedly releasing systems, finding problems and fixing them afterwards. Robinson argued that this approach is no longer sufficient as AI systems become more powerful. His criticism comes amid a wider debate in the AI industry over whether companies are developing increasingly powerful systems faster than their ability to understand and control them. OpenAI and rival Anthropic have both faced scrutiny following incidents involving failed safety controls or unexpected behaviour from experimental AI systems. Robinson also warned that AI capabilities are advancing faster than researchers’ understanding of alignment. AI alignment refers to efforts to ensure that AI systems behave in ways that are consistent with human goals and values. OpenAI defended its approach, saying it continues to put safety measures in place and can slow development when necessary. You use AI every day. Now get your AI Quotient. Take the AIQ test.