Anthropic is "blocked" by the U.S. government and has removed the Fable model; foreign media cite three major concerns: fear of aiding China in open-source AI

According to MIT Technology Review, the well-known AI startup Anthropic recently faced export restrictions imposed by the U.S. government over cybersecurity concerns related to its models "Mythos" and "Fable," forcing the company to urgently revoke access to these models. Analysts point out that this move could not only weaken America's own cybersecurity defenses but also hand the market over to China's open-source AI models, raising deep concerns about sovereignty and competition.
(Background: Micron's strategic investment in Anthropic's Series H funding! The two parties signed a long-term HBM supply agreement, $MU stock soared past 1200 and hit a new all-time high)
(Additional background: Anthropic Fable 5 caught Trump's attention—who's laughing behind the scenes? Media reveals three major beneficiaries)

Table of Contents

Toggle

• Amazon CEO as the mastermind behind the scenes? Anthropic urgently removes models
• Europe's awakening and the rise of China's open-source AI
• Weakening defenses versus Trump's policies' contradictions

The boundaries of AI safety and regulation are sparking fierce conflicts between the tech industry and the U.S. government. According to an analysis published by MIT Technology Review on June 22, 2026, the renowned AI company Anthropic recently clashed seriously with the U.S. government. This controversy involves not only AI safety but also broad discussions about national sovereignty and China's technological competition.

Amazon CEO as the mastermind behind the scenes? Anthropic urgently removes models

Reviewing the development of events, Anthropic announced in April 2026 the development of a powerful AI model called "Mythos." Due to its exceptional code processing capabilities, it was considered a potential global cybersecurity threat, so initially access was limited to a few cybersecurity experts for risk assessment. Subsequently, on June 9, Anthropic released a publicly available, security-modified version called "Fable."

However, the U.S. federal government quickly intervened that Friday, deeming the model a threat to national security and imposing export controls. Faced with government enforcement, Anthropic immediately revoked all access to Mythos and Fable. The report revealed a startling detail: it is rumored that Amazon CEO Andy Jassy informed government officials about the potential dangers of Fable. Notably, Amazon is both a major investor in Anthropic and a competitor in the AI market.

Europe's awakening and the rise of China's open-source AI

In response to this incident, author James O’Donnell highlights three key future developments to watch. First, U.S. government intervention could lead global companies to reduce reliance on American AI. French politician Bruno Retailleau called this a "wake-up call," urging Europe to accelerate building its own AI infrastructure.

However, Europe is currently facing strong challenges from China. China's open-source AI models are not only powerful and inexpensive but can also be downloaded and run directly, bypassing U.S. White House policies and protective mechanisms. This is highly attractive to companies seeking independence, even causing Chinese startup Zhipu’s stock price to surge. The article warns that whether the U.S. government will further ban domestic use of Chinese models remains uncertain.

Weakening defenses versus Trump's policies' contradictions

The second concern is that this ban could backfire. Leading cybersecurity experts have issued a public letter on freefable.org supporting Anthropic, emphasizing that their models help researchers practice defense strategies and pose no greater risk than other mainstream models. Forcibly shutting down access might actually weaken U.S. overall cybersecurity defenses, illustrating the dilemma of applying the "non-proliferation" concept to software.

Finally, this incident also puts U.S. legislators under greater regulatory pressure. It is Anthropic’s second conflict with the government (the previous involved restrictions related to Pentagon use). The article notes that although Trump once promised to relax AI safety regulations to support tech development, he has now twice listed the most valuable AI startups as national security risks. This "whimsical" policy stance makes the future of AI development highly uncertain.