Version 4

On June 12, 2026, the US Commerce Department issued an export control directive requiring Anthropic to suspend Fable 5 and Mythos 5 for foreign nationals after Amazon researchers demonstrated a classifier-based jailbreak and CEO Andy Jassy personally briefed senior Trump administration officials [3][1]. WSJ and TechCrunch reporting confirms Jassy's conversations with officials directly triggered the crackdown [4][5]. Unable to verify users' citizenship at scale, Anthropic disabled both models globally [1][2]. Anthropic is complying under legal obligation while contesting the directive as disproportionate, and emerging commentary argues the episode marks the start of government-mandated AI governance replacing company-led deployment decisions [11][12].

This is the first known use of export control authority to pull a commercial AI model after launch. The directive requires citizenship verification that no AI lab has the infrastructure to perform, making global deployment of frontier models legally uncertain under this standard. Nathan Lambert argues the episode marks the start of a new era in which technically uninformed, politically driven governance sets the rules for AI deployment — and that open-source models are likely to face similar restrictions within two years [11].

On June 12, 2026, the US Commerce Department issued an export control directive requiring Anthropic to suspend access to Fable 5 and Mythos 5 for all foreign nationals, citing a jailbreak that bypasses classifier-based safeguards for cybersecurity, chemistry, and biology prompts [1]. The directive treated providing foreign nationals access to the models as a controlled export. Anthropic determined it had no mechanism to verify users' citizenship and that disabling both models globally — including for US nationals — was the only feasible path to compliance [1][2]. Even Anthropic's own foreign-national employees who helped build the models lost access [2].

The jailbreak at the center of the directive was identified by Amazon researchers, who demonstrated they could prompt Fable 5 to produce cyberattack-assisting information it was not designed to share [3]. Amazon CEO Andy Jassy personally briefed senior Trump administration officials on those findings, and WSJ and TechCrunch reporting confirms that Jassy's conversations with officials directly triggered the crackdown [4][5]. Amazon's role is notable because it is a major Anthropic investor while simultaneously being the proximate source of the finding that prompted the restriction. White House AI and Crypto Czar David Sacks subsequently framed the directive as a request to patch a specific jailbreak with a clear path to reinstatement, not a permanent ban [6].

Anthropic published a statement complying under legal obligation while contesting the directive's technical basis: the company argues the jailbreak is minor, non-universal, and provides no capability uplift beyond what other publicly available models — including GPT-5.5 — already offer [7]. It called for a statutory review process and argued that if the recall standard were applied industry-wide, it would effectively halt all frontier model deployments [7]. Zvi Mowshowitz echoed that argument and added that restrictions driving foreign talent out of US AI labs could benefit China [6]. Critics including Semafor, Milk Road AI, and David Sacks argue that Anthropic's sustained advocacy for government AI oversight contributed to the regulatory posture it now contests [8][9][10].

Nathan Lambert, writing on June 14, offers the sharpest structural critique of all parties: he argues the episode is not an aberration but the start of a governance era in which a government that has internalized the AGI framing acts fast on technically shallow grounds [11]. He places partial blame on Anthropic itself, arguing that years of rhetoric comparing AI to nuclear weapons accelerated government willingness to take this kind of action — and warns that open-source AI advocates who are celebrating this episode are unprepared for equally aggressive restrictions on open models, which he considers likely within two years [11]. The Neuron adds a separate geopolitical concern: if the US restricts foreign-national access to its frontier models, retaliatory Chinese restrictions on frontier open-weight models could harm competitive AI pricing for US companies [12].