An OpenAI employee working on artificial intelligence safety has resigned, arguing that the company and the broader industry can no longer treat safety decisions as experiments with unlimited room for mistakes.
The departure adds to growing scrutiny of how leading AI companies manage risks as their systems become more capable and widely used.
• The former employee said safeguards must keep pace with rapid advances in AI development.
• The warning reflects concerns that commercial pressure could outstrip careful testing and oversight.
• OpenAI has faced public debate over its safety culture, governance and the balance between releasing new products and addressing potential harms.
• The resignation underscores tensions inside the industry over how much uncertainty is acceptable when deploying powerful technologies.
The employee’s comments arrive as governments and regulators push companies to provide stronger assurances on issues including misinformation, privacy, cybersecurity and autonomous decision-making. OpenAI has said safety remains central to its mission, but departures by safety-focused staff have intensified questions about whether internal processes match that commitment.