Just three days after launching Claude Fable 5, which it described as its most capable AI system yet, Anthropic has been forced to suspend access to the model after an intervention from the US government.
Earlier today, Anthropic said it received an export control directive from US authorities ordering it to block access to Claude Fable 5 and Claude Mythos 5 for all foreign nationals, including non-US employees working at the company.
The order effectively left Anthropic with little choice but to disable the models globally for all its customers, highlighting how rapidly advancing AI systems are being viewed not merely as software products but as technologies with potential national security implications.
Even as it complied with the directive, Anthropic said it has no clarity as to what exactly prompted the government’s intervention.
What Is The US Government Concerned About?
In a blog post announcing the suspension, the AI company said the directive offered few details about the national security concerns behind the move. Based on its conversations with officials, Anthropic said it believes the issue centres on a technique that could “jailbreak” Fable 5, allowing users to bypass some of the safeguards built into the model.
The company, however, disputed the significance of the finding. “… the level of capability displayed there is widely available from other models (including OpenAI’s GPT-5.5), and is used every day by the defenders who keep systems safe,” it said.
The company added that it has not even received the disclosure of the concerning non-universal potential jailbreak that led to a harmful result.
“… 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 said.
Notably, this is not the first time that Anthropic has come under the US administration’s scrutiny. Earlier this year, the Pentagon reportedly designated the company an “unacceptable supply chain risk”. The latest directive goes a step further by restricting access to the models based on nationality.
What Are Fable 5, Mythos 5?
The latest development came days after Anthropic announced the launch of Fable 5 and Mythos 5, describing them as a major step forward in AI capabilities.
Introduced on June 9, Fable 5 is designed as a general-purpose version of Anthropic’s new Mythos-class models, a category the company says surpasses its Opus family in areas ranging from software engineering and scientific research to visual reasoning and complex problem-solving.
Alongside Fable 5, Anthropic introduced Mythos 5, a version of the model with some cybersecurity safeguards lifted. Rather than opening it to new users, the company initially made it available only to organisations that already had access to Claude Mythos Preview under Project Glasswing, its cybersecurity-focused programme run in collaboration with the US government.
It is pertinent to mention that the launch of Claude Mythos Preview had raised concerns across countries on its potential misuse. Following that, Anthropic expanded Project Glasswing earlier this month to around 150 organisations across more than 15 countries, including India, enabling participants to identify and remediate critical software vulnerabilities before they could be exploited by attackers.
Indian Startup Ecosystem Calls For National AI Strategy
The latest developments have reignited a long-running debate over India’s dependence on foreign AI platforms and the need for a more effective national AI strategy.
It is pertinent to note that the Centre, under the ₹10,372 Cr IndiaAI Mission, is offering subsidised AI compute, startup financing support, among other AI ecosystem initiatives. However, investors and founders urged the government to do more following the US administration’s move.
Zoho founder Sridhar Vembu said the development underscores the growing link between technology and national power. “Technology is the ultimate weapon. National sovereignty, national security, all of it is now about technology,” he said, adding that “globalisation is dead and Bharat must find her own way ahead”.
Lightspeed Venture Partners’ Hemant Mohapatra described the development as a “sovereign AI is real” moment.
Mohandas Pai, chairman of Aarin Capital, called for a large-scale national AI push and proposed a dedicated annual fund of ₹50,000 Cr for deeptech and AI, arguing that India must accelerate investments in strategic technologies.
The development is also expected to lead to startup founders rethinking how AI products are built. Saurabh Awasthi, cofounder of Augmen.IO, said the suspension is a reminder that startups cannot afford to become overly dependent on a single frontier AI provider.
“This is an alarm — for startups and nations alike,” he said. “AI models being treated as strategic assets is a reminder to avoid over-dependence on frontier models.”
He argued that startups need to build resilience through multi-model architectures, open-source fallback models and on-premise deployments.
However, Shayak Mazumder, founder and CEO of Adya.ai, described the suspension as a “knee-jerk reaction” by the US government and said it is likely temporary. He argued that while Fable 5 is a strong model for long-horizon tasks, its advantages over leading models such as GPT-5.5, Gemini and Anthropic’s Opus series are often overstated.
“I don’t see this being a call to action that India needs its own large foundational models,” he said. “Instead, we should spend time and money building smaller models, creating fine-tuning studios, and creating the world’s best harnessed models.”
The post Anthropic’s Fable 5, Mythos 5 Access Suspension Rekindles Sovereign AI Debate appeared first on Inc42 Media.
https://ift.tt/mdIvZf1
0 Comments