President Donald Trump's administration is asserting direct control over the rollout of OpenAI’s newest flagship model, GPT-5.6 Sol, marking an escalation in executive branch involvement with artificial intelligence. OpenAI announced Friday that the model would launch “in the coming weeks,” but only after the government signs off on which enterprises may use it. The company described Sol as its most powerful model yet.
The move follows a similar action earlier in June, when the administration effectively banned Anthropic's advanced “frontier” model shortly after its release. That model—originally called “Mythos” and later released in a trimmed version named “Fable”—reportedly contained a potential security flaw. Together, these steps suggest the White House is building a de facto pre-approval system for cutting-edge AI.
The White House's stance has drawn sharp partisan lines. Republicans in Congress have largely supported tighter executive oversight of frontier models, arguing it protects national security. Democratic lawmakers and civil liberties groups have expressed concern that such gatekeeping could let the administration pick winners and losers in the AI industry, stifling competition and innovation.
Public opinion on the matter remains divided. Recent polls indicate that while a majority of Americans worry about AI safety, a sizable minority—including many tech workers—fears government intervention could hinder progress. The policy's electoral implications could ripple through the 2026 midterms, with swing-state voters in tech-heavy districts likely to be closely watched.
Legal experts note that the administration's approach marks a significant departure from the traditional research-and-approval structure used during prior administrations. If challenged in court, the government would likely need to demonstrate a clear national-security rationale for picking individual companies for access, a standard that has yet to be tested.