The Trump administration has requested OpenAI to delay the release of its GPT-5.6 model, citing potential security concerns. OpenAI will comply by releasing the model in a limited preview form for a small group of enterprise customers, who will need to get approval on a case-by-case basis. This deal is more favorable than one given to rival Anthropic, which was forced to suspend access to its models due to export control directives. Why it matters: This move highlights the administration’s willingness to regulate AI in a way that favors American companies, raising concerns about the tech industry’s freedom to innovate.