① AI startup Anthropic has shut down access to its cutting-edge models Fable 5 and Mythos 5 in response to a US government export control order, sparking ongoing controversy.
② Anthropic customer Legion has sued the US government, claiming the ban has caused irreparable harm to its business.
③ Some experts are calling for the revocation of the control order, arguing it may ultimately benefit US adversaries.
June 24 — The controversy continues over AI startup Anthropic’s decision to shut down access to its cutting-edge models Fable 5 and Mythos 5 due to a US government export control order. On Tuesday, a startup that uses Anthropic’s models filed a lawsuit against the US government.
The lawsuit comes less than two weeks after Anthropic complied with the US government ban by closing access to the models. The plaintiff, Legion, is a litigation technology company that develops tools for lawyers. Headquartered in the United States, Legion states that its software developers include Canadian citizens working in Canada.
In its filing, Legion claims that Anthropic’s closure of access to the Fable 5 model, in particular, means the company instantly lost the latest tool at the core of its development. The directive has caused immediate, irreparable, and existential harm to the company.
The defendants named in the lawsuit include Commerce Secretary Lutnick, who warned Anthropic’s CEO in a letter that the company would need US government authorization before exporting the Fable 5 and Mythos 5 products outside the United States or providing them to foreign nationals anywhere.
The rule has also raised market concerns that closed-source models are vulnerable to political interference, leading to increased attention on open-source models.
Controversy
The Mythos model has drawn international attention since its release earlier this year. Its powerful performance in discovering cybersecurity vulnerabilities has led to restricted access to the model, while Fable 5 is a simplified version of Mythos.
A US official said Tuesday that Anthropic has worked with US intelligence agencies to use its Mythos model for testing and identified certain vulnerabilities within hours, but this does not mean the model can exploit those vulnerabilities in a short period.
Although Anthropic has recently collaborated with US agencies to test vulnerabilities in its AI systems, tensions between the company and the Trump administration are escalating, with Anthropic expressing concerns about how the US military uses its AI systems.
However, a spokesperson for Anthropic said in a statement that Anthropic appreciates the government’s efforts to address security concerns regarding the Mythos model.
The company will continue to work with the government to protect critical infrastructure and ensure the United States maintains its leadership in artificial intelligence.
Some artificial intelligence experts and cybersecurity executives argue that the Trump administration should rescind the directive, as such a move may benefit U.S. adversaries more than it harms them.
More than 100 cybersecurity experts and corporate leaders, including those from Adobe and Nvidia, stated in a letter to the government that Anthropic’s Mythos model is highly effective at discovering software flaws and weaponizing vulnerabilities, but it is not uniquely dominant.
They added that stripping the United States of its best cyber defense capabilities amid intense international competition is a dangerous course of action.
