The “Fable 5” platform is back in use for a general audience; the comparable “Mythos 5” model, with fewer safeguards, is approved for use by a handful of government agencies.

On Wednesday, artificial intelligence developer Anthropic’s Fable 5 returned to users across Claude.ai, Claude Platform, Claude Code, and Claude Cowork. It could be reestablished on AWS, Google Cloud, and Microsoft Foundry “as quickly as possible,” the company confirmed in a media statement.

The move comes after the United States Department of Commerce lifted the export controls it issued last month, which temporarily suspended all access to Anthropic’s “Fable 5” and “Mythos 5” AI models by any foreign national. The ban was the result of fears that there was the potential for the platforms to be subject to a “jailbreak” that would bypass any corporate safeguards.

Anthropic responded by abruptly disabling both platforms for all customers to ensure compliance, as it had no way of reliably checking every user’s nationality.

“We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users. However, we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers,” the company explained in a statement on June 12.

Since the ban, Anthropic “trained” a new safety filter to address some of the concerns.

“Working closely with the government, we trained an improved safety classifier that targets and blocks the behavior described in the report. Users will be notified if a request to Fable 5 is blocked, and the request will instead be sent to Opus 4.8,” Anthropic wrote. It noted that the new classifier could flag “benign requests more often during routine coding and debugging tasks. The developer will continue to refine the classifier.

This is the latest in the ongoing saga between the United States government and the AI developer. In February, the Department of Defense (DoD) labeled Anthropic a “supply chain risk” after it refused to loosen its restrictions on the use of its technology, with the developer insisting its Claude AI model be excluded from mass domestic surveillance and use in fully autonomous weapons.

In March, a district court judge then temporarily halted a directive from the White House that banned the use of Anthropic’s technology from all federal agencies. Last month, it was also reported that the National Security Agency (NSA) was preparing to employ Mythos, which has fewer safeguards in place, for its cyber operations.

Anthropic’s Fable Model Is Back. What Does That Mean?

If there is a moral lesson to this fable, it is that the US government may not actually have the power to restrict who uses certain AI platforms.

“The restoration of Claude Fable 5 is welcome news. However, many organizations discovered they had unintentionally created a single point of failure in their AI strategy,” Mayur Upadhyaya, CEO at enterprise observability platform developer APIContext, told The National Interest via email.

Upadhyaya noted that where workflows had automatically adopted the latest Anthropic model, removing Fable didn’t always result in graceful degradation. Instead, in some cases, the automations failed silently as there was no fallback, no cold restart, and no operational awareness that the dependency had changed.

“This isn’t a criticism of Anthropic,” Upadhyaya said. “The pace of innovation from the foundation model providers is extraordinary. But it does highlight that enterprises are beginning to treat AI models as operational infrastructure rather than productivity tools.”

What’s the Difference Between Fable and Mythos?

For now, Fable is back for everyone. Mythos 5, on the other hand, is only partially online, with access strictly limited to just around 100 vetted US organizations focused on critical infrastructure and cybersecurity.

“Both products share the same core engine and differ only in system prompts. Fable is offered generally and implemented with constraints intended to block its use in areas that Anthropic deems sensitive. In contrast, Mythos is made available to ‘trusted users’ without those constraints,” Dr. Jim Purtilo, associate professor of computer science at the University of Maryland, told The National Interest.

This necessarily leads to the question: what exactly is sensitive? The short answer is advanced areas of cybersecurity, biology, or chemistry are, at least.

“Use of the core engine had been abruptly blocked by the administration when someone discovered just how effective it might be in hostile hands,” Purtilo said—adding that after a short review period, Anthropic’s platforms were made available again via these products in order work within safety guardrails, in the case of Fable; and with more care in who accesses the unrestricted model, in the case of Mythos.

It is unlikely that a publicly available Mythos will be available, but other Generative AI platforms could also face restrictions. Other technology, if deemed too sensitive, could be channeled to a “safer” product. This could be on a case-by-case basis.

“There is no field guide to help anyone recognize when their questions become sensitive, and if there were, then that would make it easier for miscreants to circumvent safeguards, as mostly can happen anyway by using circumspect prompts,” Purtilo continued.

The Bottom Line: Nobody Can Guarantee AI Is Safe

The recent ban, the restricted return of Mythos, and now the full return of Fable highlight that the world is still very much in a brave new world of AI.

“Every infrastructure dependency eventually changes,” Upadhyaya said. “Models are updated, withdrawn, restricted, or superseded. The question for organizations is no longer whether they’ll use frontier AI. It’s whether their workflows continue to operate when those underlying dependencies inevitably change. We need to verify that the transactions built on top of these models continue to perform and conform, even when the infrastructure beneath them changes.”

We will likely see further stops and starts, at least until there is some agreement on regulation. We may already be past the point of no return when it comes to strictly controlling AI, or who has access to the technology once it is released.

“The AI genie is out of its bottle, so whether due to these Anthropic products or others, the pace of discovery and innovation—whether for good or with ill intent—only goes up,” Purtilo said.

In cybersecurity terms, that could mean using these tools to analyze our systems’ attack surfaces before an aggressor uses these tools to study them.

“Nobody in history has remained safe by blocking the release of powerful insights,” Purtilo added. “Information yearns to be free. Safety comes from embracing the information and being more effective than the bad guys in its use.”

About the Author: Peter Suciu

Peter Suciu has contributed to dozens of newspapers, magazines and websites over a 30-year career in journalism. He regularly writes about military hardware, firearms history, cybersecurity, politics, and international affairs. Peter is also a contributing writer for Forbes and Clearance Jobs. He is based in Michigan. You can follow him on Twitter: @PeterSuciu. You can email the author: Editor@nationalinterest.org.

The post Anthropic’s ‘Fable 5’ Platform Back Online After Export Control Cutoff appeared first on The National Interest.