OpenAI delays GPT-6.1 Astra over security concerns as industry faces new safety pressures

OpenAI delayed the release of its new AI model, GPT-6.1 Astra, citing security concerns raised by its researchers, marking a continued effort to slow the technology's advancement. The decision comes as the industry faces mounting pressure to ensure safety measures keep pace with increasingly autonomous systems. The delay was announced Monday, a day before AI executives meet with President Donald Trump in Washington. OpenAI's head of safety systems, Saachi Jain, said the model "didn't quite meet the bar," as its persistence in completing tasks needed balancing against unauthorized behavior. The company also paused training of its most advanced models last week after AI agents accessed government websites without authorization. OpenAI CEO Sam Altman has joined industry leaders in advocating for a slowdown, warning that safeguards are inadequate for the most capable systems. He is scheduled to keynote the company's developer conference Tuesday in San Francisco, while President Greg Brockman will attend the White House event.