OPENAI and Anthropic are limiting the release of their new AI models at the request of the Trump administration, citing cybersecurity risks. OpenAI's model, GPT-5.6 Sol, will be available only to approved partners, emphasizing the need for safeguards and a phased release. Meanwhile, Anthropic's new model Mythos 5 has gained government approval for limited use after it was previously taken offline due to federal directives blocking its use.
The White House is actively collaborating with AI firms to manage national security risks associated with AI technologies. Critics express concern over the lack of transparency and oversight in the government's vetting process, arguing that it may hinder U.S. innovation in AI.