The administration under President Donald Trump has implemented a prohibition on foreign access to premier AI models developed by Anthropic, citing urgent national security exigencies. This decision highlights the U.S. government’s ongoing strategy regarding export controls associated with advanced technologies.
This recent decree from the United States arrives shortly after Anthropic, the entity responsible for the Claude chatbot, unveiled its latest AI offerings – the Claude Fable 5 and Mythos 5 models.
“The U.S. government, invoking national security protocols, has issued a directive halting access to Fable 5 and Mythos 5 for any foreign nationals, whether located domestically or internationally, inclusive of foreign national employees of Anthropic,” the company articulated in an official blog post.
This maneuver has rekindled tensions between Anthropic and the Trump administration. The San Francisco-based firm is currently embroiled in litigation against the government due to its placement on a supply chain blacklist for resisting military requests to deploy its AI systems for domestic surveillance and fully autonomous weaponry.
Will this latest directive from the U.S. exacerbate the rift between the AI company and the White House?
Underlying Reasons for the U.S. Order
According to Anthropic, the United States government communicated its concerns regarding national security but failed to elucidate further specifics.
China has made notable advancements in AI technology recently, exemplified by the launch of generative AI models DeepSeek-V3 and DeepSeek-R1, developed by the Chinese enterprise DeepSeek at significantly reduced costs compared to their American counterparts.
Beijing’s monopolistic hold on rare earth supply chains, crucial to the AI industry, further exacerbates these concerns.
David Sacks, an advisor to President Trump, revealed in a post on X that the government became aware of potential security vulnerabilities associated with Anthropic’s AI model, prompting the imposition of the export control order.
Sacks indicated that there were warnings suggesting Fable 5 could be compromised, and stated that when Anthropic’s co-founder and CEO, Dario Amodei, was informed, the company took no corrective actions.
In its blog entry, Anthropic asserted that the U.S. government had only provided “verbal evidence” concerning a potential narrow, non-universal jailbreak of its model.
“We contest the notion that the discovery of a narrow theoretical jailbreak justifies the withdrawal of a commercial model already utilized by millions,” the company stated.
Kirsten Davies, the Pentagon’s chief information officer, expressed on X on June 13 that the Defense Department is firmly committed to prioritizing national security.
“Some priorities transcend revenue cycles, clickbait, and pre-IPO valuations. America First. Always,” Davies remarked.
However, Al Jazeera has been unable to independently verify the accuracy of these media reports.
Impacted AI Models
The Fable 5 model, launched by Anthropic earlier this week and grounded in Mythos technology, has been expressly prohibited from access due to the U.S. government’s mandate.
Experts have apprised Reuters that the Mythos models, if misused, possess the potential to drastically enhance sophisticated cyberattacks, particularly in industries like banking that depend on intricate, interconnected, and often obsolete technological frameworks.
Anthropic stated this week that it has collaborated with the U.S. government and other entities on safety issues preceding the Fable launch, asserting that rival AI systems also exhibit capabilities to identify minor coding bugs.
Other prominent AI developers, including OpenAI, have produced similar models but have typically circumvented such stringent restrictions imposed by the Trump administration.
In response to the latest governmental mandate, Anthropic remarked on Friday: “The net effect is that we must suddenly disable Fable 5 and Mythos 5 for our clients to maintain compliance. Access to all other Anthropic models will remain unimpeded.”
Impact of the U.S. Order on Global Research and Foreign Workers
The recent directives delineate the boundary of software access produced in the U.S. and are aligned with the Trump administration’s policy of imposing export controls on vital technologies, particularly in the realms of AI and semiconductors. Previous limitations have also been enacted regarding chip sales by tech titans like Nvidia and AMD.
This directive may severely undermine global research and development, as research entities, including international collaborators, are poised to lose access to pioneering technological innovations.
At present, numerous global firms, including the equity and research organization S&P, utilize Anthropic’s Claude software to streamline their databases, thus enabling financial advisers, investors, and analysts to access S&P data more efficiently.
Moreover, foreign nationals working in the U.S. on H1-B visas, along with individuals residing outside the U.S., are now barred from accessing Anthropic’s advanced AI models.
Several key personnel at Anthropic, including co-founder Chris Olah, AI researcher Andrej Karpathy, and philosopher Amanda Askell, were born outside the United States. However, it remains uncertain whether they will also face restrictions regarding software access.
The ramifications are likely to disrupt workflows and productivity for U.S. companies that rely on Anthropic’s services, with numerous employees rendered unable to utilize the firm’s sophisticated software.
Some users on X have articulated additional implications of this order.
Kun Chen, a notable member of the tech community and former L8 engineer at Meta, Microsoft, and Atlassian, asserted that the U.S. government has erred with this directive.
“Employing ‘foreign national’ as the criterion for model access is not only impractical but leads to extensive prohibitions.
Moreover, it offers little utility, as numerous Americans could be more adversarial towards the U.S. than foreign nationals, and the method is easily evaded by those with malicious intent,” he elucidated.
Sridhar Vembu, co-founder of the Indian multinational enterprise Zoho, commented on X that this order underscores “technology as the paramount weapon.”
“National sovereignty and security now revolve around technology,” he emphasized.
Vembu further articulated that this presents an opportunity for nations like India to bolster their own advancements in AI.
In the meantime, Anthropic has indicated that it perceives a “misunderstanding” regarding the situation and is actively working to restore access to its models expeditiously.
Source link: Aljazeera.com.
