Tech news in 3 minutes
The White House is asking OpenAI to slow roll the release of its new model over safety concerns
OpenAI’s upcoming GPT 5.6 model will not be released to the public like previous versions. Instead, the company plans to share it only with a select group of close partners, following pressure from the Trump administration, according to The Information. In a meeting this week, CEO Sam Altman reportedly told staff that the government would be “approving access customer by customer” during a preview period. If the limited release goes well, OpenAI hopes to follow with a broader public release “a couple of weeks later.” The Trump administration appears to be pushing OpenAI to keep its most powerful AI models under wraps, similar to what Anthropic is already doing voluntarily. According to The Information, OpenAI’s new model is under administration review, and its staff “worked closely” with the government on the upcoming release. The agencies that requested a limited release were the Office of the National Cyber Director and the Office of Science and Technology Policy. The administration, which initially favored a “hands-off” approach to AI, has recently pushed for federal oversight. Earlier this month, Trump signed an executive order directing certain AI companies to voluntarily submit new models for government testing and evaluation before public release. Earlier this year, Anthropic sparked controversy by releasing its Claude Mythos model only to a small group of partners via Project Glasswing, arguing the model was too powerful and could cause harm in the wrong hands. Observers debate whether this is a marketing gimmick or a genuine safety measure. Cybercriminals have long used automated tools, but generative AI has increased threats. LLMs can write malware and even autonomously execute ransomware. Frontier cyber tools like Mythos can identify and exploit software vulnerabilities at speeds unmatched by human analysts, posing significant risks for organizations with complex software infrastructure. However, since these models remain closed to the public, their actual threat level is difficult to assess.