**
In a surprising turn of events, Anthropic, a leading name in artificial intelligence, has temporarily disabled access to its highly anticipated AI models, Fable 5 and Mythos 5. This move comes in response to a US government export control directive aimed at safeguarding national security. The company describes the situation as a “misunderstanding” and is eager to restore access to its technology as swiftly as possible.
A Sudden Withdrawal
On Friday, Anthropic announced that it had no choice but to “abruptly disable” its flagship AI offerings for all users. The directive was issued by the US government, which cited national security concerns without providing extensive details. The underlying issue appears to be a potential method for bypassing safeguards that prevent Fable 5 from identifying software vulnerabilities—an alarming prospect for many in the tech community.
This development arrives at a time when relations between Anthropic and the US government were showing signs of improvement. Earlier in the year, tensions flared when Anthropic chose not to allow the US military to employ its AI for domestic surveillance and autonomous weaponry purposes. Consequently, the company found itself on a government blacklist, a move that is set to take effect later this year.
The Broader Implications
The latest directive signals a significant escalation in the US’s efforts to restrict the capabilities of foreign adversaries in the AI domain. Historically, US export controls have concentrated on the hardware and tools that facilitate AI development, rather than directly limiting access to the AI models themselves. This shift could reshape how AI is developed and shared internationally.
Anthropic has voiced its disagreement with the government’s stance, asserting that the evidence provided was merely verbal and not substantial enough to warrant such a drastic action. “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 stated.
Interestingly, just days prior to the directive, Anthropic had been a proponent of increased US governmental oversight of AI technologies, advocating for the ability to block models deemed too risky. However, the company now claims that the recent government action does not align with principles of fair and evidence-based regulation.
The Response from Officials
Kirsten Davies, the Pentagon’s chief information officer, expressed her support for the government’s prioritisation of national security via a post on social media platform X. “Some things are simply more important than revenue cycles, clickbait, and pre-IPO valuation. America First. Always,” she affirmed, highlighting the administration’s commitment to protecting national interests over corporate gains.
Earlier in the week, Anthropic had just unveiled Claude Fable 5, an advanced model equipped with guardrails designed to prevent misuse in sensitive areas like cybersecurity. Critics, however, have noted that these guardrails may be excessively restrictive, potentially hindering innovation while also opening the door to misuse that could expedite sophisticated cyberattacks in vulnerable sectors such as banking.
The Road Ahead
In response to the government’s directive, Anthropic has made it clear that it is committed to compliance but views the situation as a misunderstanding. The company is working diligently to rectify the access issue, stating, “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”
Amazon Web Services (AWS), which hosts Anthropic’s models, confirmed that the company had requested to revoke access to Fable 5 and Mythos 5 for all users in all regions. A US official corroborated that the Commerce Department’s directive effectively suspends access for foreign nationals, raising questions about the implications for non-American staff at Anthropic, some of whom are integral to its operations.
Why it Matters
The fallout from this directive encapsulates the growing friction between AI developers and regulatory bodies as the landscape of technology continues to evolve. As governments grapple with the potential risks associated with advanced AI, the implications extend beyond just one company. This situation could establish precedents that redefine how AI models are developed, shared, and governed on a global scale. The outcome will not only affect Anthropic but could also have lasting effects on the future of AI innovation and international collaboration in this rapidly advancing field.
