The US government is exercising control over the release of advanced AI models worldwide, forcing OpenAI, Anthropic, and other leading AI companies to delay or halt the release of new models
Situation Summary
The Trump administration is using export control regulations to restrict the dissemination of advanced AI models deemed critical to national security, creating an unprecedented situation in the AI industry.
OpenAI Delays GPT-5.6 Launch Following Government Request
OpenAI, which was planning to release GPT-5.6 this week, had to announce a delay in its launch after receiving a request from the US government, primarily due to "national security concerns" and preventing the use of AI technology for potentially dangerous purposes. GPT-5.6 is classified as a highly capable model capable of handling complex multi-level tasks, which the government believes requires control to prevent it from being used to develop weapons or activities dangerous to national security.
Anthropic Ordered to Shut Down Mythos 5 and Fable 5 Globally
In an even more severe case, the US government ordered Anthropic to shut down Mythos 5 and Claude Fable 5 models worldwide without exception. This order sent shockwaves through the tech industry as it was the first time the government exercised power to shut down already operational AI models.
Changes After 2 Weeks of Tension
Following two weeks of tension between the government and AI companies, the US government recently allowed Anthropic to release Mythos 5 in a limited form to certain companies and government agencies. This relaxation is part of "Project Glasswing," a special program permitting the use of advanced AI models in specially approved groups with strict usage control conditions.
Impact on the Global AI Industry
1. Changes in AI Regulation
This marks a significant turning point showing that AI is not just ordinary technology but a national security issue requiring strict control.
2. Impact on AI Development in the US
AI companies in the US must be cautious in developing and releasing new models, as they may be controlled or banned from release at any time.
3. Commercial Interests
These controls may put US AI companies at a disadvantage in the global market when companies in other countries can develop and release models freely.
Expert Perspectives
AI policy experts view these controls as a "double-edged sword" with both positive and negative aspects.
Positive aspects:
Preventing AI from being used for dangerous purposes Creating safety standards for advanced AI
Negative aspects:
Putting the US at a disadvantage in global AI competition Potentially creating mistrust between government and technology companies
Future of AI Regulation
These events suggest that in the future we may see:
Clearer AI regulation laws
Specialized AI regulatory agencies
Restricted AI technology sharing between countries
Conclusion
The US government's control of AI models in June 2026 represents a historical turning point showing that AI has become a matter of national security importance. The government's ability to order the halt or delay of leading AI model launches demonstrates the "soft power" of the state in controlling the most powerful technology of this era. What will be interesting to watch is how the AI industry adapts and develops under tighter regulatory frameworks.
References
- The Guardian - OpenAI Delays GPT-5.6 Launch
- CNBC - US Government Allows Anthropic Limited Release of Mythos 5
- The New York Times - US Relaxes Restrictions on Anthropic's Mythos AI Model
- 9to5Mac - US Government Orders Global Shutdown of Claude Fable 5 and Mythos 5