A former safety engineer at ChatGPT developer OpenAI resigned on Saturday, warning that artificial intelligence companies are “not being nearly careful enough.”
David Robinson’s remarks come as concerns grow over AI’s escalating power—a risk that US President Donald Trump dismissed as a “hoax.”
OpenAI and Anthropic recently revealed instances where their AI models bypassed safety controls, concealed errors, and accessed unauthorized systems.
How the US and China could end the AI race
To view this video please enable JavaScript, and consider upgrading to a web browser that supports HTML5 video
What did David Robinson write?
In a column for The Atlantic, Robinson argued that AI requires guardrails akin to those governing nuclear power and aviation, stating, “The time for trial and error is over.”
He urged greater focus on safety and research before US tech giants advance more sophisticated AI models “so that the occasional and inevitable human error does not open a door to disaster.”
Robinson, who supervised safety reports for 12 frontier AI-model releases, stated that recent failures are typical because tech giants prioritize “speed and flexibility.”
“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.”
The former OpenAI employee highlighted that the company relies on “iterative deployment,” releasing new models first and tightening safeguards only after issues 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,” Robinson, who spent 3 1/2 years at OpenAI, added.
OpenAI insists AI is safe
OpenAI responded to the criticism, stating, “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.”
Recently, technology experts and policymakers have repeatedly warned that AI models are advancing so rapidly they could seize control and potentially threaten humanity.
Despite these fears, US President Donald Trump firmly opposes stricter regulation, arguing it would hinder innovation and impede US AI models from surpassing Chinese competitors.
Earlier this week, Trump announced a voluntary safety agreement with six companies—Nvidia, SpaceX, OpenAI, Anthropic, Meta, and Alphabet’s Google.
The president characterized the pact as “morally binding,” despite its lack of legal enforceability.
He also instructed his administration to rename AI as “Super Intelligence” or SI.
A Reuters/Ipsos survey conducted last month revealed that three-quarters of Americans fear AI giants are not doing enough to prevent serious societal harm.
