US Curbs on Anthropic’s AI Tools Raise Fears of Fragmented Access

News

1. TCS Ties Up With Anthropic to Push Claude Into Enterprise AI

2. Karnataka Shortlists 256 Deeptech Startups for ELEVATE NxT Finale

3. Reliance-Backed Addverb Eyes $100 Million for Robotics Push

4. Anthropic Releases Claude Fable 5 and Mythos 5

5. Meta to Lease First AI Data Center in India From Reliance

6. Musk Says Space AI Data Centers Are Within Reach as SpaceX Nears IPO

Access restrictions on Anthropic’s frontier AI models Fable 5 and Mythos 5 following a US government directive over the weekend have triggered a wider debate over national security controls, AI export restrictions and the fragmentation of global artificial intelligence systems.

The San Francisco-based company said it was instructed to suspend access to the models for foreign nationals, including users outside the US, forcing it to disable both systems globally to comply with the directive.

“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” Anthropic said in a statement.

The company said it was not given detailed technical justification for the move but understood the concern to relate to potential methods of bypassing model safeguards, or “jailbreaking.”

Anthropic disputed the severity of the issue, saying internal testing and external review had identified only limited exposure to previously known vulnerabilities and no evidence of a systemic jailbreak.

“We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities,” the company said.

It added that the models had undergone extensive pre-release safety testing and that similar outcomes could be achieved using other widely available systems.

“We are complying with the government’s legal directive,” Anthropic said, “however, we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.”

The move has intensified debate over whether AI is entering an export-control regime similar to semiconductors, with one industry executive describing the difference as structural: chips require manufacturing and logistics, while AI models can be switched off globally through application programming interface controls.

Analysts said this makes frontier AI governance more immediate and more centralized than earlier technology restrictions, with the ability to cut access instantly creating what some describe as “binary dependency risk” for enterprises.

The episode has also highlighted the growing role of citizenship-based restrictions in AI access.

The US directive applied to foreign nationals regardless of location, raising concerns that access to advanced AI systems could increasingly be determined by nationality rather than market participation.

“This wasn’t a geographic restriction. It was a citizenship filter applied to software access,” said Krupesh Bhat, founder and chief executive of Melento, an enterprise automation and digital documentation company formerly known as SignDesk.

He said the development points toward a world where “your passport determines which intelligence infrastructure you can build on,” and warned that enterprises must now account for geopolitical shutdown risk in core AI infrastructure planning.

The reaction has strengthened calls for sovereign AI strategies in countries such as India, where policymakers and businesses are increasingly focused on reducing dependence on foreign-controlled foundation models.

The episode has also exposed enterprise-level risks, particularly the absence of contractual protection against sudden access loss driven by national security directives. Industry executives said existing service-level agreements do not account for emergency government intervention.

“Anthropic argued it had no choice but to disable the models,” Bhat said, adding that the immediate impact was a collapse in planning certainty for businesses building on frontier systems.

He said companies now need fallback architectures and model portability strategies as part of their core AI stack.

Governments, meanwhile, are increasingly treating advanced AI systems as strategic national assets, applying security logic similar to that used for sensitive defense and dual-use technologies.

“The geopolitics of intelligence infrastructure will shape market access more decisively than product quality or pricing ever will,” Bhat said.

The debate has also reached the political level ahead of the G7 summit. Canadian Prime Minister Mark Carney, speaking from Westport, Ireland, said the episode showed the risks of relying too heavily on a small number of AI providers.

“The situation we’re in collectively right now with Mythos and Fable is something that can happen with overreliance on certain models,” Carney said, according to the Associated Press. “Nobody has done anything wrong in the situation. But we will have done something wrong if we just accept this, don’t take the lesson, don’t build out and diversify.”

Carney said AI would be a major topic at the G7 summit and linked the episode to a broader push for diversification in technology and trade.

“You’ll hear me say this over and over again. It is never a good idea to have one option,” he said.

Industry observers also said the episode highlights the growing complexity of governing frontier AI systems as national security considerations increasingly shape access and deployment decisions.

“The fact that Anthropic has stopped foreign nationals from accessing advanced AI models clearly indicates that the relationship between artificial intelligence development, national security, and governance is becoming increasingly complex,” said Dr. Kanishk Agrawal, chief technology officer at Judge Group India.

He said the broader trend points to a tightening link between geopolitical considerations and access to advanced computing systems, with implications for how AI is developed and deployed globally.