OpenAI has announced it will delay the launch of its next artificial intelligence model, GPT-5.6, and restrict its access at the request of the White House. This comes as the Trump administration has recently intensified security measures for AI companies, including export controls on Anthropic's AI models.
According to reports from U.S. tech outlets such as The Information and CNN on June 25, the White House's National Cybersecurity Agency and Office of Science and Technology Policy requested OpenAI to postpone the launch of GPT-5.6 for security assessments. Additionally, the White House has instructed OpenAI to allow only a limited number of approved institutions to access the model prior to its full release.
Sam Altman, CEO of OpenAI, communicated this decision during an internal Q&A session, stating that the company would delay the launch and limit access to the model. He told employees, "We have made it clear to the U.S. government that this is not our preferred long-term model, and we will work with government and industry stakeholders to establish a more sustainable approach for future releases." As a result, the launch of GPT-5.6, initially scheduled for this week, is now expected to be pushed back to July.
Axios noted that this marks the first instance of the U.S. government proactively intervening in the release of an AI model by an American company. Altman reportedly discussed the GPT-5.6 launch with Secretary of Commerce Howard Lutnick, who expressed a desire for government agencies to thoroughly test the model for security concerns before its release.
The Trump administration has recently blocked foreign access to Anthropic's AI models, Mythos5 and Fable5, due to security concerns, and plans to ensure that future AI models undergo adequate security evaluations before being launched. According to a source cited by CNN, the administration views the performance of OpenAI's upcoming GPT-5.6 as being on par with Mythos.
However, The Verge reported that the measures taken against OpenAI are considered less severe than those imposed on Anthropic. Earlier this month, President Trump signed an executive order requiring AI companies to voluntarily request government reviews 30 days before launching new AI models. Nonetheless, CNN pointed out that a framework for implementing this executive order has yet to be established, leading to confusion among AI companies regarding oversight authorities. Notably, the guidance for OpenAI came from the White House, while the directives for Anthropic were issued by the Department of Commerce.
As competition intensifies between U.S. and Chinese AI companies, Axios noted that the Trump administration's control measures are placing additional burdens on American AI firms.
Meanwhile, The New York Times reported that OpenAI may delay its planned initial public offering (IPO) until next year due to stock price pressures following the recent IPO of SpaceX.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.