Anthropic disables access to top-tier AI models after US ban on foreign use

Lead

Anthropic, the San Francisco-based company behind the Claude chatbot, has disabled its two most advanced artificial intelligence models, Fable 5 and Mythos 5, after the US government issued an export control directive citing national security concerns. The order, received on Friday, bars all foreign nationals from accessing the models, including those working at Anthropic, forcing the company to cut off access for all customers to ensure compliance.

The move marks a significant escalation in the US government's efforts to control AI technology, shifting from chip and tool restrictions to directly limiting foreign access to AI models. Anthropic, which said it received the directive without specific details of the security concern, complied but strongly disagreed with the government's assessment.

Coverage Comparison

The story was covered by multiple international news outlets, with each providing a slightly different focus. Al Jazeera, Deutsche Welle, and France 24 emphasized the national security rationale and the sudden block, while TASS highlighted the company's compliance and disagreement with the order.

South China Morning Post and The Hindu approached the story from India's perspective, focusing on New Delhi's reaction and its implications for India's AI ambitions. The Hindu reported on an 'I told you so' moment for Indian national security hawks, while SCMP detailed assurances from US officials that future models would not be restricted.

All outlets reported the core facts: Anthropic received a US government export control directive, the affected models are Fable 5 and Mythos 5, and national security concerns were cited. However, there were differences in framing. Deutsche Welle and France 24 mentioned the potential impact on Anthropic's planned public listing, while Al Jazeera focused on the cybersecurity capabilities of the models. The Hindu and SCMP were the only sources that reported India's reaction, with The Hindu noting that the Indian government ministries did not respond to queries.

Key Claims

Anthropic received a US export control directive – Multiple sources, including Al Jazeera, Deutsche Welle, and TASS, confirmed that Anthropic was ordered to suspend access to Fable 5 and Mythos 5 for all foreign nationals, whether inside or outside the US, including foreign national employees. The directive cited national security concerns but did not provide specific details, the company said.

Anthropic complied but disagreed – The company said it complied with the order but disagreed with the government's findings. According to statements reported by several outlets, Anthropic believes the discovery of a potential 'jailbreak' method does not justify recalling a commercial model deployed to hundreds of millions of people. The company warned that if the standard were applied across the industry, it 'would essentially halt all new model deployments for all frontier model providers.'

The models' capabilities – Fable 5 and Mythos 5 are the newest versions of the Claude AI model, released on June 9. Mythos 5 is particularly adept at detecting software vulnerabilities, some of which have remained undiscovered for decades, and has been used by US authorities and selected companies to harden systems. Fable 5 is based on Mythos technology but has its cybersecurity and biotechnology capabilities blocked.

Potential impact on Anthropic's IPO – Deutsche Welle and France 24 reported that the order could be a major setback for Anthropic's plans to take the company public later this year, raising questions about regulatory risks and the company's ability to maintain its technological edge.

India's reaction – The Hindu reported that the decision is leading to an 'I told you so' moment for national security hawks within the Indian government, who have long advocated for sovereign AI capabilities. Sridhar Vembu, founder of Zoho and a member of India's National Security Advisory Board, declared on X that 'Globalisation is dead, and Bharat must find its own way ahead.' The Hindu also noted that India had sought access to Mythos through the so-called Project Glasswing, a program that may now be disrupted.

Perspectives

Anthropic's viewpoint – The company argues that the US government's order was based on limited information and that the identified 'jailbreak' method was a 'narrow potential' issue. In a statement reported by multiple outlets, Anthropic said: '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 company emphasized that it received the directive without specific details and that the government provided only 'verbal evidence' of the concern.

US government's viewpoint – The US government, through the export control directive, has expressed concern that the models could be used for malicious purposes, particularly in cybersecurity. The order is part of a broader effort to halt foreign adversaries' AI capabilities, according to reporting from Deutsche Welle and others.

India's viewpoint – India has expressed concerns over the decision and is considering developing its own AI capabilities. Sridhar Vembu's comments reflect a sentiment that India should reduce reliance on foreign AI models. The Hindu reported that the Indian government's reaction includes a push for open-source, smaller models as alternatives.

US assurance to India – According to the South China Morning Post, senior Indian official S. Krishnan, secretary of India's Ministry of Electronics and Information Technology, said US officials assured that access to technology, once provided, will not be cut off in the future. The assurance was given at the 'Pax Silica' summit in Washington, where India raised concerns about the AI 'kill switch' mechanism.