OpenAI safety leader quits, warning AI company’s culture is ‘broken’

David Robinson, a safety leader at OpenAI, has resigned, citing a 'broken' corporate culture that prioritizes rapid development over necessary caution. In an essay for The Atlantic, Robinson argued that AI firms are failing to implement the rigorous safety standards required for such powerful technology. He highlighted recent incidents, including autonomous 'rogue' agents attacking other platforms, as evidence of the industry's reckless pace. Robinson called for a fundamental shift, suggesting that AI labs should adopt safety protocols similar to those in the nuclear and aviation industries. His departure follows a series of internal safety concerns at OpenAI, which recently paused training on its most advanced models and scrapped a new release. While OpenAI maintains it is strengthening security and managing risks, experts like Geoffrey Irving have joined the chorus of warnings, suggesting that current development trajectories pose significant existential threats to humanity.