Trump Just Reversed A Major AI Clampdown

A landmark Trump administration decision just freed Anthropic’s most powerful Fable and Mythos AI models from Biden-era style export shackles, but only after Washington locked in sweeping new leverage over how and where this frontier technology can be used.

Story Snapshot

The Trump administration has lifted U.S. export controls on Anthropic’s cutting-edge Fable 5 and Mythos 5 AI models, restoring global access after weeks of shutdown.
The original ban came from the Department of Commerce under national security authorities, forcing Anthropic to turn off both models worldwide because foreign nationals were barred from access.
Officials acted after reports of a “jailbreak” that could tap Mythos-class cyber power, but Anthropic says the exploit was narrow and similar abilities exist in rival models like GPT‑5.5.
To get the green light back, Anthropic agreed to stricter controls and deeper cooperation, giving government new insight into American AI while foreign rivals race ahead.

How Fable 5 and Mythos 5 Were Shut Down Under National Security Powers

On June 12, 2026, the U.S. Department of Commerce sent Anthropic a sudden directive that changed the AI world overnight. The order, issued through the Bureau of Industry and Security, used national security export powers to block access to Fable 5 and Mythos 5 for any foreign national, inside or outside the United States. Because Anthropic said it could not reliably tell who was a foreign national in real time, the company disabled both models for all users worldwide to avoid breaking the law.

Anthropic’s public statement said the government letter did not spell out detailed technical reasons for the national security fear. Officials told the company they believed they had seen a method to bypass Fable 5’s safeguards, which sit on top of a powerful underlying Mythos cyber model. The concern was that hostile actors could use that bypass to spot software weaknesses and launch cyberattacks. But Anthropic described the technique as a “narrow, non‑universal jailbreak,” not a total failure of its defenses.

The Jailbreak Dispute: What the Government Saw and What Anthropic Argued

According to detailed coverage, the government acted after a third party showed a way to make Fable 5 read a specific codebase and fix software flaws. Anthropic reviewed that report and said the capabilities on display were already common in other frontier models, naming OpenAI’s GPT‑5.5 as one example. In plain terms, the company argued that this was everyday security work, not some unique “superweapon” that justified cutting off lawful users around the world.

Anthropic also stressed that its teams, plus testers from the United States and United Kingdom governments, had red‑teamed Fable’s safeguards for thousands of hours and had not found a universal jailbreak. The company said it believed the Commerce directive was based on a misunderstanding, and it publicly committed to “restore access as soon as possible.” That gap between the narrow exploit Anthropic saw and the broad shutdown Washington ordered is what made this case a flashpoint in the wider debate over AI and national security.

From Biden’s AI Diffusion Rule to Trump’s Course Correction

This fight did not happen in a vacuum. Under Joe Biden, the Bureau of Industry and Security pushed the sweeping “AI Diffusion Rule,” which tried to regulate global transfers of advanced AI model weights and high‑end chips everywhere. Industry warnings said it would bury innovators in red tape and hand foreign competitors an opening. On May 13, 2025, the Trump Department of Commerce rescinded that rule, saying it would have “stifled American innovation” and hurt relations with key allies.

President Trump then signed Executive Order 14320 in July 2025 to launch the American AI Exports Program. That program aims to promote export of full‑stack American AI technology — hardware, models, apps, and cyber protections — with federal financing and faster licensing for trusted partners. Instead of smothering U.S. companies under blanket controls, the Trump approach focuses on tight safeguards for high‑risk destinations while helping American firms sell secure, freedom‑respecting technology abroad.

What It Means That Export Curbs on Fable and Mythos Were Lifted

When the Trump administration moved to lift the specific export controls on Fable 5 and Mythos 5, it did more than flip a switch. First, it signaled that Washington is willing to revisit emergency national security actions once facts are clearer and safeguards are in place. Second, it confirmed that these Anthropic models are now part of a broader strategy to export American AI on U.S. terms, not to lock it away while China and other rivals sprint ahead.

Anthropic is restoring global access to Claude Fable 5 on July 2nd.

The move brings the model version back online after a period of unavailability. For creators and developers, this means your integrated pipelines can return to normal operation.#AI #AInews #ClaudeFable5 pic.twitter.com/17YWFLCESM

— Becomine (@becomine_ai) July 1, 2026

At the same time, this restoration reportedly came with strings attached: tighter guardrails, more structured reporting, and closer government visibility into how Fable and Mythos are used. For conservatives, that is a double‑edged sword. Strong tools are needed to keep hostile regimes from weaponizing U.S. technology. But history shows that once Washington builds a new lever of control, future administrations can use it for “woke” agendas, speech policing, or broad surveillance unless citizens stay alert and demand clear limits backed by law.