100+ Experts Urge Trump to Lift Export Controls on Anthropic’s 2 AI Models

The Trump administration’s abrupt restriction on Anthropic’s Fable 5 and Mythos 5 AI models raises fundamental questions about the limits of executive authority over digital exports. More than 100 cybersecurity and legal experts are now calling for an open, scientific assessment process, arguing that the opaque directive may set a dangerous precedent for AI governance without congressional oversight or stakeholder consultation.

The Trump administration has imposed what appears to be an unprecedented restriction on artificial intelligence models, blocking foreign nationals from accessing Anthropic’s latest frontier systems, Fable 5 and Mythos 5. The company complied on Friday June 12, 2026, taking the models offline. Within days, a coalition of over 100 cybersecurity executives and experts—including representatives from Adobe and Nvidia—sent a letter urging the White House to rescind the export controls, warning that the move could backfire by ceding defensive capabilities to America’s adversaries, particularly China. The letter argues that China’s own models are “only months behind,” and that advanced domestic open-source counterparts already provide similar vulnerability-hunting abilities, making the restriction more harmful than helpful. This clash illustrates the deepening tension between national security and open innovation in the AI era, raising questions about the legal and strategic framework governing frontier AI.

“

The Trump administration’s abrupt restriction on Anthropic’s Fable 5 and Mythos 5 AI models raises fundamental questions about the limits of executive authority over digital exports.

Anthropic’s Fable 5 and Mythos 5 models are described as capable of surpassing human cybersecurity experts in discovering and exploiting software vulnerabilities. The company had previously limited access to select customers due to these capabilities and had engaged with the White House regarding the risks. However, the administration’s directive—the most significant step yet to restrict the most advanced AI models—abruptly severed access for all foreign nationals, forcing Anthropic to take the models offline entirely. The company has publicly stated it does not believe the government’s steps were warranted by the security concerns it initially flagged. This disconnect suggests a breakdown in the collaborative risk-assessment process that both industry and government had been attempting to build.

The open letter from 100+ cybersecurity leaders, including execs from Adobe and Nvidia, strikes at the core of the debate. The signatories contend that while Mythos 5 is “quite good” at finding and weaponizing software flaws, it is “not uniquely good.” Many of them already use other foundation and open-source models for security audits and training. Thus, they argue, removing the best defensive tools from US cybersecurity professionals (who are themselves foreign nationals or operating in global contexts) without a clear reason is dangerous because it hampers the very defenders who need to outpace state-backed adversaries. The letter explicitly notes that China’s models are only months behind the best American models, and likely possess private capabilities beyond what’s publicly known—making the US self-imposed disadvantage a strategic blunder.

The regulatory action signals an intensification of export controls on AI beyond hardware (such as semiconductors) to include software and model weights. This represents a novel application of trade statutes and national security authorities, likely under the International Emergency Economic Powers Act (IEEPA) or similar mechanisms. Legal experts will scrutinize whether the administration’s action followed proper rulemaking procedures and consultation with the scientific community. The letter’s demand for “an open, scientific and transparent process of handling AI risk assessments” hints at frustration over opaque decision-making that bypassed industry input. The incident may set a precedent for future AI model restrictions, potentially triggering a wave of legal challenges from affected companies and civil liberties groups concerned about overreach.