An OpenAI safety employee has quit and is sounding the alarm

David Robinson, an OpenAI safety employee whose work included writing the safety reports that accompanied every major model release at the company, resigned from his position this week, according to The Verge. He is now speaking out in an editorial published in The Atlantic.

According to the report, Robinson says the culture in the industry is fundamentally broken. He characterizes the problem as deeper than simply adding a few new rules or regulations governing how training models are handled.

The Verge notes that readers may feel cynical about people suddenly coming forward to warn about the dangers of technology they helped build, and acknowledges that they contributed to the situation. The report adds that this does not mean their warnings should be discounted.

The excerpt does not include further specifics from Robinson's editorial, such as particular incidents or recommendations. No statement from OpenAI is included in the available material.