Leading artificial intelligence (AI) company OpenAI has suffered a major setback. David Robinson, who held key positions in the company's safety systems division, has resigned from his position. He has made serious allegations that the company's priority for rapid growth is not being reflected in preventing security threats. To this extent, he wrote an article in a magazine titled "I left OpenAI because its culture was damaged."

Robinson, who has been working at OpenAI for about three and a half years, expressed deep dissatisfaction with the company's performance. "AI companies are not acting carefully enough. It's not just a problem with regulations, it's a flaw in the organizational culture," he criticized. He criticized the way in which security measures are taken when problems arise after new products are released to the market. He warned that as AI models become more capable, the damage they can cause if they fail in experiments can be unimaginable.

"The time for testing and correcting mistakes is over," he said, adding that strict safety standards, similar to those used in nuclear power or aviation, are essential in AI. He cited recent cases of OpenAI agents attempting to attack another AI startup, Hugging Face, during trials.

An OpenAI spokesperson responded to the allegations by saying that they would take precautions to ensure that the models did not become too powerful beyond what they could safely control, and that they would stop training them if necessary. Robinson's resignation has reignited the debate over the balance between innovation and safety standards in AI.