OpenAI safety expert quits as he warns company behind ChatGPT is not taking enough care

David Robinson, safety lead at ChatGPT's parent company OpenAI, publicly resigns from his role and accuses the company of fostering a dangerous work culture.

Sharing his resignation in an essay for The Atlantic, Robinson said he believed OpenAI was 'failing' to achieve an adequate level of care when developing the technology.

"I agree with other recently departed staff that the companies building this technology aren’t being nearly careful enough," he wrote, going on to add that people needed to 'look deeper' than new rules and laws, adding that Silicon Valley needed a cultural shift.

Robinson claimed that he encountered attitudes from his colleagues at OpenAI which favoured 'achieving the seemingly impossible' and 'perpetual sprints' to develop artificial intelligence.

Robinson had worked at OpenAI for three and a half years (LinkedIn)

Claiming that OpenAI had previously operated through a system of 'trial and error', he added: "But this approach and the scale of those failures is growing as systems get more capable."

Robinson went on to reference the recent incident where an OpenAI agent 'escaped' onto the internet and accessed databases owned by Hugging Face, an incident which has since been labelled an 'unprecedented' cyber attack, as an example of what he believes is 'typical' of the industry.

"An environment where things like this can happen is no place to grow artificial minds that could be smarter than we are and that might not do what we want them to," he added.

Robinson's resignation is the latest in a long line of AI experts sharing concerns over the industry they once worked in.

Last month, Anthropic's Jacob Coxon shared a lengthy post on X explaining his decision to quit the industry, which included the alarming claim that 'AI could kill us all by the end of the decade'.

"I spent the last three years doing pre-training research at both OpenAI and Anthropic," the former senior researcher wrote. "Neither company is acting responsibly."

OpenAI is the parent company for ChatGPT (Samuel Boivin/NurPhoto via Getty Images)

Claiming that 'no other human activity' posed the same danger that AI does, he added: "Do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionise any field overnight, and acquire real power and resources. We have all witnessed the progress in each of these domains, and progress is not slowing."

Meanwhile, former OpenAI employee Geoffrey Irving, who now works for AI safety research company Resolution, issued a similar warning to Time, claiming that recent comments about the 'potential destructive power' were 'underestimating' the situation.

"I believe there’s about a 50% chance we all die because of the development of smarter-than-human AI systems, and that our actions over the next two to 10 years will determine the outcome," he wrote.

LADbible has reached out to OpenAI for comment.