By Isqil Najim
The global technology ecosystem was thrown into a tailspin over the weekend following a sweeping directive by the United States Department of Commerce, which slapped strict national security export controls on Anthropic’s latest artificial intelligence models, Fable 5 and Mythos 5.
The development, which has triggered intense conversation among tech analysts and policy experts globally, effectively bars the distribution of these advanced models to individuals outside the US, as well as foreign nationals living within American borders—including Anthropic’s own non-citizen workforce. Facing an enforcement bottleneck due to the sheer scope of the order, Anthropic announced it had been left with no choice but to completely disable the models for all users worldwide.
According to Anthropic, the regulatory hammer was pulled after research on a “jailbreak”—a term used to describe bypassing an AI’s built-in ethical and security safeguards—sparked panic within Washington’s policy corridors.
However, reacting to the development, Katie Moussouris, the Chief Executive Officer of Luta Security, a prominent cybersecurity firm, has heavily faulted the US government’s rationale. Speaking with the Wall Street Journal, Moussouris revealed she had reviewed the research paper in question, which was compiled by researchers at Amazon using specific prompts to identify system vulnerabilities.
‘An Own Goal’ Against National Defense
“I’ve seen the paper. It’s not a jailbreak,” Moussouris clarified in a detailed post on the social media platform X (formerly Twitter) on Saturday. She argued that the researchers were merely engaged in Defense Oriented Prompting (DOP)—a critical technical capability that cybersecurity defenders require to fortify digital infrastructure.
In a sharp rebuke of the Joe Biden-led administration’s knee-jerk reaction, the cybersecurity boss added: “If National defense is the goal, this just scored an own goal against us.”
Anthropic has also kicked against the sudden regulatory crackdown. In an official blog post, the frontier AI firm expressed strong reservations about the government’s approach, arguing that punishing an organization for discovering a narrow vulnerability sets a dangerous precedent for the entire tech industry.
“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,” Anthropic stated. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”
While conceding that the American government possesses the legal mandate to restrict unsafe AI technologies, Anthropic insisted that such interventions must follow a statutory, transparent, and fair process anchored on verifiable technical facts. “This action does not adhere to those principles,” the company lamented.
Tech Balkanisation and the Cost of Transparency
Meanwhile, the inclusion of foreign nationals inside the US in the ban has raised severe red flags among global tech policy analysts.
Ben Murphy, a scholar at the Institute for Progress think tank, described the directive as “another step on the balkanization of technology”—a term referring to the fragmentation of the global internet and tech space into isolated regional pockets based on geopolitical lines.
“It might previously have been unthinkable to require proof-of-citizenship to access services; it’s increasingly common across new technologies and, to be honest, this attempt is not surprising in that light,” Murphy noted on X.
Policy experts are also warning that Washington’s aggressive stance could yield unintended negative consequences for the AI ecosystem. Analysts fear that AI laboratories may become highly secretive, keeping advanced models completely in-house or pulling them from public access entirely to avoid regulatory traps.
Furthermore, observers point out that Anthropic’s commitment to corporate transparency appears to have backfired spectacularly. The company was already locked in a running battle with the US government, which previously designated it a supply-chain risk for Pentagon contractors. Despite the frosty relationship, Anthropic had voluntarily granted the government early access to its Mythos model while proactively flagging its potential cybersecurity implications.
“I don’t know that the government wouldn’t have reached that conclusion themselves, but as a business matter, those pronouncements have not produced a healthy working relationship with the government,” Murphy added, highlighting the growing trust deficit between Silicon Valley and Washington DC.
As of press time, the US Department of Commerce had not responded to media inquiries regarding the sudden enforcement. On its part, e-commerce and cloud giant Amazon—a major stakeholder and cloud provider—chose to play its cards close to its chest.
“As a leading cloud provider that serves a large number of private and public sector customers, it’s not uncommon for governments to seek our counsel on potential security risks. When they occur, we don’t share the details of these discussions,” an Amazon spokesperson told Fortune.