AI Security Scrutiny
OpenAI and Anthropic Restrict New AI Models Amidst Government Cybersecurity Review for Select Partners
OpenAI has restricted its newest AI model, GPT-5.6 Sol, to customers vetted and approved by the Trump administration, Post Register reported. The move is part of an unprecedented government review to check AI products for cybersecurity risks before they reach the public.
Anthropic, maker of the Claude chatbot, faces similar limits. The Commerce Department effectively banned Anthropic's most powerful cybersecurity model, forcing the company to restrict access to a small group of trusted partners, according to Barchart.
GPT-5.6 Sol Locked Behind White House Approval
OpenAI's GPT-5.6 Sol is one of its most capable models to date. But instead of a broad public launch, the company agreed to release it only to customers the Trump administration has cleared, Yakima Herald reported. This is the first time a sitting U.S. government has directly shaped who can access a cutting-edge AI model.
Both OpenAI and Anthropic announced their newest models would go first to small groups of trusted partners. The White House has been working closely with frontier AI labs — companies building the most advanced AI systems — to scale up the technology while managing risks, according to WFMZ.
Commerce Department Bans Anthropic's Top Cybersecurity Model
Anthropic took a harder hit than OpenAI. The Commerce Department effectively banned the company's strongest cybersecurity-focused AI model from wide release, Lancaster Online reported. Anthropic then chose to limit access on its own rather than fight the order.
The ban marks a major shift in how the U.S. government treats AI tools built for cybersecurity. Such tools can be used to find — or exploit — weaknesses in computer systems. Officials appear worried about who might get access to that power, according to Rutland Herald.
Dario Amodei and Anthropic Draw Blame for Stricter Rules
Some in the industry point fingers at Anthropic CEO Dario Amodei for the tighter government scrutiny. Critics say his public warnings about AI dangers gave regulators reason to step in, according to Erie News Now. Amodei has long argued that powerful AI needs strong safety controls.
Others argue Anthropic's own choices — building highly capable cybersecurity tools — made government review inevitable. Either way, the result is the same: two of the biggest AI labs in the world now need White House sign-off to reach their customers, HNG News reported.
What This Means for the Future of AI Releases
The vetting process sets a new standard for how AI models get released in the United States. If the government can approve or block access to a model, it holds real power over which companies succeed and which customers get the best tools, according to Voice of Alexandria.
The White House has said it wants to lead the world in AI development. But critics warn that restricting access could slow progress and give rivals in other countries an edge. For now, OpenAI and Anthropic are playing by the administration's rules, Lufkin Daily News reported.
Sentiment
NeutralPublishers
18
Articles
17
Reach
18
