Last Friday evening, Anthropic’s most powerful publicly available AI model, Claude Fable 5, suddenly disappeared.
One minute, people were testing its AI superpowers; the next, they were getting error messages. I know because I was one of them.
Within hours, Fable 5 and its sister model, Mythos 5, had been taken offline worldwide following an emergency directive from the United States Government.
Antropic posted their statement on X.
The reason? National security concerns over a reported jailbreak technique that officials believed could allow the model to assist with offensive cyber activities.
Anthropic disputes the severity of that assessment, but complied with the order nonetheless.
This seems to be the first government intervention of its kind with a frontier AI model, and it’s led to many conversations this week about its possible future impact.
Will governments take control of AI models? What does this mean for business and our use of Frontier models if we become heavily reliant on them?
Many organisations are building with AI models and integrating systems such as Claude into their daily workflows.
Just days earlier, Anthropic had launched Fable 5 as its most capable public model to date. And I got on board. I worked with Fable 5 on my business plan and forecasting the direction of my company, given a few things I am working on.
Anthropic itself described Fable 5 as being safety controlled:
“We have instituted strong safeguards that greatly reduce the likelihood that Fable is misused for tasks related to cybersecurity (among others). In fact, our safeguards are so strong that many users have complained that they are overly broad.”
The company disagreed with the assessment and stated that it believed there had been a misunderstanding, but it complied with the directive while discussions continued, and in order to comply, removed access for everyone, not just non-US citizens as per the US government order.
Anthropic could not immediately separate every US user from every non-US user across its customer base.
So it chose the simplest option available. It switched both models off. For everyone.
Now, regardless of where you sit on the politics of this decision, I think it marks an important moment for AI.
For years, we’ve talked about AI regulation in theory. And especially around the EU AI Act.
I’ve discussed safety frameworks. We’ve heard calls for stronger oversight of frontier models. Last Friday, we saw what oversight looks like in practice. A government intervened, and a frontier AI model disappeared.
Anthropic’s position is that vulnerabilities should be identified, patched and improved through ongoing safety engineering, much like cybersecurity flaws are managed in software.
Many people think about AI risk in terms of hallucinations, bias, copyright, or data privacy. All of those issues matter. But Fable 5 highlights a different category of risk altogether. Dependency risk.
If your organisation relies heavily on a single frontier model, what happens when access changes overnight?
AI models are becoming part of a much larger ecosystem involving governments, regulators, national security agencies, and technology companies.
Access may increasingly depend on geography, citizenship, industry sector, or regulatory status.
So, if the AI system your organisation relies on disappeared tomorrow, what would happen next? Would the work stop? Would key processes break? Would your team have alternatives?
That’s the conversation I think leaders need to start having. Because AI is no longer just a technology story. It’s a business story. It’s a governance story. And increasingly, it’s a geopolitical story.
If you want to build your AI literacy and skills, I can help
📚 Grab AI Playbook on Amazon, the perfect podcast companion.
📰 Subscribe to Human in the Loop, my LinkedIn Newsletter
🎓 Take the CPD Accredited AI Playbook Diploma
📞 Book a call at digitaltraining.ie
Sources
- Anthropic Statement: US government directive to suspend access to Fable 5 and Mythos 5
- Reuters: Anthropic disables top-tier AI models after US order limiting foreign access
- Business Insider: Inside the whirlwind 24 hours that led the White House to slap export controls on Anthropic
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
