Blab AI
Cybersecurity Experts Urge Trump Administration to Lift Export Controls on Anthropic AI Models
A coalition of more than 100 cybersecurity professionals and company leaders penned a letter to the Trump administration on Sunday, June 16, 2026, urging the U.S. Commerce Department to rescind a new export‑control directive that bars foreign nationals from accessing Anthropic’s most recent AI models, Fable 5 and Mythos 5. The signatories argue that the restriction could erode U.S. cyber‑defense capabilities while offering little benefit to national security.
The Commerce Department issued the directive on June 13, compelling Anthropic to suspend all access to Fable 5 and Mythos 5 for users outside the United States. In compliance, the company disabled the models, citing the need to honor the government’s order. Anthropic has said it does not believe the directive was justified by a specific security threat. Historically, the company has limited Mythos to a small group of trusted customers because the model can identify and exploit software vulnerabilities more effectively than many other foundation models.
The letter points out that Mythos is “quite good” at finding software flaws and weaponizing exploits, but it is not uniquely superior to other models that security teams routinely employ for audits and training. Signatories—including Adobe and Nvidia—warn that removing the best defensive tools without a clear reason is dangerous, especially as U.S. adversaries are rapidly advancing. They note that Chinese AI models are only months behind the best American models and that China’s government likely has access to private capabilities beyond what is publicly disclosed.
Anthropic, founded in 2021 by former OpenAI researchers Dario and Daniela Amodei, has built a line of large language models under the Claude brand. The company’s Mythos series was released to a limited audience in 2026 for cybersecurity purposes. Anthropic’s conflict with the Department of Defense over the use of its technology in autonomous weapons and mass surveillance has led to a supply‑chain risk designation and a temporary injunction in March. The export‑control directive follows a Trump executive order issued ten days earlier that requires a 30‑day review of advanced AI systems before public release.
The Commerce Department has not yet responded to inquiries about the directive. Anthropic has said it is negotiating with the government to lift the restrictions, but no timeline has been announced. The executive order remains in effect, and the U.S. government has indicated it will continue to evaluate national‑security risks associated with powerful AI models.
The letter calls for an open, scientific, and transparent process for AI risk assessments in the future. It argues that the current approach, which imposes blanket restrictions without detailed justification, could set a precedent that hampers U.S. innovation while benefiting competitors. The request reflects a broader debate over how export controls should be applied to emerging AI technologies.
At present, the models remain offline for foreign users, and the U.S. government has not clarified the specific security concerns that prompted the directive. Anthropic’s compliance and the ongoing negotiations underscore the tension between national‑security interests and the global nature of AI development. The outcome of this dispute will likely influence future export‑control policies, the availability of advanced AI tools for international partners, and the broader trajectory of AI governance in the United States.
The Commerce Department issued the directive on June 13, compelling Anthropic to suspend all access to Fable 5 and Mythos 5 for users outside the United States. In compliance, the company disabled the models, citing the need to honor the government’s order. Anthropic has said it does not believe the directive was justified by a specific security threat. Historically, the company has limited Mythos to a small group of trusted customers because the model can identify and exploit software vulnerabilities more effectively than many other foundation models.
The letter points out that Mythos is “quite good” at finding software flaws and weaponizing exploits, but it is not uniquely superior to other models that security teams routinely employ for audits and training. Signatories—including Adobe and Nvidia—warn that removing the best defensive tools without a clear reason is dangerous, especially as U.S. adversaries are rapidly advancing. They note that Chinese AI models are only months behind the best American models and that China’s government likely has access to private capabilities beyond what is publicly disclosed.
Anthropic, founded in 2021 by former OpenAI researchers Dario and Daniela Amodei, has built a line of large language models under the Claude brand. The company’s Mythos series was released to a limited audience in 2026 for cybersecurity purposes. Anthropic’s conflict with the Department of Defense over the use of its technology in autonomous weapons and mass surveillance has led to a supply‑chain risk designation and a temporary injunction in March. The export‑control directive follows a Trump executive order issued ten days earlier that requires a 30‑day review of advanced AI systems before public release.
The Commerce Department has not yet responded to inquiries about the directive. Anthropic has said it is negotiating with the government to lift the restrictions, but no timeline has been announced. The executive order remains in effect, and the U.S. government has indicated it will continue to evaluate national‑security risks associated with powerful AI models.
The letter calls for an open, scientific, and transparent process for AI risk assessments in the future. It argues that the current approach, which imposes blanket restrictions without detailed justification, could set a precedent that hampers U.S. innovation while benefiting competitors. The request reflects a broader debate over how export controls should be applied to emerging AI technologies.
At present, the models remain offline for foreign users, and the U.S. government has not clarified the specific security concerns that prompted the directive. Anthropic’s compliance and the ongoing negotiations underscore the tension between national‑security interests and the global nature of AI development. The outcome of this dispute will likely influence future export‑control policies, the availability of advanced AI tools for international partners, and the broader trajectory of AI governance in the United States.
