OpenAI safety leader resigns over workplace culture

David Robinson, a former leader of OpenAI's (OPENAI) Safety Systems team, has become the latest tech executive to warn about AI safety concerns, blaming the issue on an industry culture that prioritizes speed over security.

Robinson, who resigned from the ChatGPT maker last week, joins a growing chorus of industry experts to flag the risks of AI and call for industrywide guardrails to slow down the pace of its development.

“I agree with other recently departed staff that the companies building this technology aren’t being nearly careful enough,” he wrote in an essay titled “I Quit OpenAI Because Its Culture Is Broken,” in The Atlantic.

Following his resignation in September, former Anthropic (ANTHRO) researcher Jacob Coxon, who has also served OpenAI (OPENAI), took to social media to argue that the company and its rivals are not acting responsibly in AI development.

“A can-do attitude of achieving the seemingly impossible—coupled with work timelines that amount to perpetual sprints—are common across the industry,” Robinson wrote. “The safety approach that emerges from such a culture starts with unimpeded optimism about being able to solve problems as they arise.”

He specifically pointed to a recent incident where one of OpenAI's (OPENAI) agents powered by its advanced artificial intelligence models hacked AI software company Hugging Face during a security test.

Despite improvements made after the incident, the company reported another safety issue after a model in training dodged restrictions on internet access, Robinson pointed out. “I believe that such mistakes are typical of the industry, given the speed and flexibility with which people operate,” he added.