Anthropic sends top engineers to White House to lift restrictions on its most powerful AI models

Anthropic has sent senior engineers to Washington to resolve a dispute with the White House that led to restrictions on its most advanced AI models. The episode has also sparked calls from Zoho's Sridhar Vembu for India to strengthen its own AI capabilities.

In Short

• Anthropic engineers head to Washington over AI dispute
• White House concerns led to Mythos 5, Fable 5 restrictions
• Sridhar Vembu calls for stronger Indian AI efforts

Anthropic has reportedly dispatched senior technical staff to Washington as the AI company works to resolve a growing dispute with the Trump administration that resulted in restrictions on some of its most advanced AI models. According to a fresh report by Axios, Anthropic engineers have been holding discussions with White House officials in an effort to address concerns that led to the company's flagship AI models, Claude Mythos 5 and Claude Fable 5, being taken offline. Sources familiar with the matter told the publication that both sides are interested in finding a resolution and restoring normal operations.

The development comes just a few days after Anthropic was forced to disable access to the two models following a federal export control order. The order reportedly required the company to prevent non-US nationals from accessing the models, leading to a sudden shutdown of services that had only recently been made available to paying Claude subscribers.

Why Anthropic's AI models were taken offline

The restrictions stem from concerns raised by the US government over the capabilities and security of Anthropic's latest AI systems. Mythos 5 and Fable 5 belong to the company's "Mythos-class" family of models, which are built on the same technology foundation as Claude Mythos Preview. Anthropic had previously described that earlier model as too powerful for public release because of the potential risks associated with misuse.

While Mythos 5 and Fable 5 were introduced with additional safeguards intended to reduce those risks, concerns reportedly emerged shortly after launch. According to earlier reports, several companies, including Amazon, informed White House officials that they had identified ways to bypass some of the models' protections through so-called jailbreak techniques.

The concerns appear to have escalated rapidly. Reports suggest that White House officials contacted Anthropic on Friday and warned that keeping the models online posed a national security risk. The company was allegedly given a short window to respond before a formal export control order was delivered. Within hours, access to the models was shut down.

Another report from Semafor claimed that US officials were also examining concerns that a group linked to China may have gained access to a Mythos-class model. Anthropic, however, does not officially allow access to its AI systems from China.

The latest Axios report suggests the administration was unhappy with Anthropic's initial response to complaints about potential vulnerabilities. That dissatisfaction may have contributed to the unusually aggressive action taken against the company.

The situation has quickly become one of the biggest flashpoints in the AI industry, raising questions about how governments will regulate increasingly powerful artificial intelligence systems. It also highlights the growing national security concerns surrounding advanced AI models, particularly those considered capable of handling sensitive cybersecurity-related tasks.

Sridhar Vembu urges India to reduce dependence on foreign AI providers

The White House action has also triggered reactions from technology leaders outside the US. Zoho founder and chief scientist Sridhar Vembu called the development a major sign that access to advanced technology is increasingly becoming a matter of national security.

In a post on X, Vembu wrote, "This is big: all access to Mythos and Fable AI models disabled for everyone outside America."

He argued that countries can no longer assume unrestricted access to the world's most advanced technologies. According to Vembu, India should focus on building its own AI capabilities, encourage wider use of Indian and open-source AI models, and deepen domestic research efforts. He also pointed out that training frontier AI models requires enormous computing resources and access to advanced chips, both of which are becoming harder to obtain due to growing geopolitical restrictions.

Vembu said the latest developments show how quickly technology and national interests are becoming intertwined. He urged Indian organisations to reduce dependence on foreign AI providers and invest in homegrown alternatives that can continue to operate regardless of global policy changes.

- Ends