What's happened
Anthropic has resumed access to Claude Fable 5 and Mythos 5 for select U.S. organizations after export controls were lifted, while OpenAI is restricting GPT-5.6 Sol at the government’s request. The moves follow a June crackdown over safety concerns and possible jailbreaking methods.
What's behind the headline?
Analysis
- The tempo of AI regulation is accelerating as governments move from blocking to phased reintroduction, signaling a shift from risk avoidance to risk management.
- The staff-level nature of the controls suggests a broader, ongoing framework rather than a single rule; more publishers may face similar limits as they expand access.
- The practical impact will be felt by businesses relying on advanced AI for cybersecurity tasks, with access now granted to select entities under government oversight.
What’s driving the update
- Authorities are balancing innovation with security, aiming to curb misuse while permitting legitimate use by vetted groups.
- The reintroduction hints at confidence in safeguards, even as lawmakers push for tighter oversight and clearer standards.
reader takeaway
- Expect more controlled reopenings and stricter licensing for high-risk AI tools over the coming weeks.
How we got here
Export controls were imposed on June 12 and are being lifted in a staggered manner. Anthropic and OpenAI have cited government concerns about vulnerabilities and potential weaponization. The developments come as authorities continue to vet advanced AI systems for national security.
Our analysis
AP News reports Anthropic has publicly stated Fable 5 is widely available again and Mythos 5 is accessible to select U.S.-based organizations under federal approval. Independent and BBC corroborate the timing of export-control changes and provide context about the jailbreaking concerns previously raised by authorities. OpenAI has also announced temporary, government-ordered restrictions on GPT-5.6 Sol.
Go deeper
- Which AI models are most affected by export controls right now?
- How might these controls affect your organization’s use of AI tools?
- What safeguards are currently in place for high-risk AI systems?
More on these topics
-
United States - Country in North America
The United States of America, commonly known as the United States or America, is a country mostly located in central North America, between Canada and Mexico.
-
OpenAI - Artificial intelligence company
OpenAI is an artificial intelligence research laboratory consisting of the for-profit corporation OpenAI LP and its parent company, the non-profit OpenAI Inc.
-
Anthropic - Artificial intelligence company
Anthropic PBC is a U.S.-based artificial intelligence startup public-benefit company, founded in 2021. It researches and develops AI to "study their safety properties at the technological frontier" and use this research to deploy safe, reliable models for
-
Amazon - E-commerce company
Amazon.com, Inc. is an American multinational technology company which focuses on e-commerce, cloud computing, digital streaming, and artificial intelligence.