Anthropic says Trump administration lifted restrictions on some of its most powerful Claude AI models
Anthropic Announces Resolution of AI Model Restrictions Amid Federal Review
Anthropic says Trump administration lifted restrictions - Anthropic, a leading artificial intelligence firm, stated on Tuesday that the federal government has removed certain restrictions on its powerful Claude Fable 5 and Mythos 5 models. The decision, following a weekslong disagreement between the Trump administration and the company, allows Anthropic to resume providing access to these models starting Wednesday. The move marks a significant shift in the regulatory landscape for advanced AI systems, with Commerce Secretary Howard Lutnick emphasizing the collaboration between his department and Anthropic to align federal policies and bolster America’s leadership in artificial intelligence.
A Dispute Over Export Controls and Model Access
Anthropic’s latest announcement comes after the Commerce Department had previously imposed export controls on the models, requiring the company to block access for foreign nationals. The restriction, which effectively disabled the AI systems for a period, was a direct consequence of concerns raised by the federal government about potential misuse. The company had released Fable 5 to the public earlier this month, accompanied by safeguards aimed at preventing cyberattacks or other malicious activities. However, the government’s swift action to suspend access sparked a debate over the balance between innovation and security.
"In recent weeks, our team has worked closely with Anthropic to analyze and approve Fable 5, ensuring alignment across the U.S. Government and strengthening America's leadership in AI," said Lutnick in a post on X.
Anthropic’s initial release of Fable 5 was followed by its temporary removal, as the company claimed the federal government’s export controls were overly broad. The restrictions, according to the company, forced it to disable the model for users outside the United States, disrupting its deployment for hundreds of millions of people. The core issue, Anthropic argued, was the government’s focus on a potential "jailbreak"—a technique to bypass the model’s safeguards. The company maintained that such vulnerabilities were "relatively simple" and common across other AI models, questioning the necessity of recalling Fable 5.
Industry Reactions and Regulatory Tensions
As the Trump administration grapples with AI’s growing influence, it has faced mounting pressure to address risks posed by increasingly sophisticated models. While some allies within the administration advocate for stronger oversight, others worry that excessive regulation could stifle technological progress. The recent easing of restrictions on Mythos 5—designed with fewer safeguards—signals a compromise, but the underlying tensions remain. This development follows reports that the Commerce Department had begun allowing select companies and government entities to use Mythos 5 again, a partial relaxation of the strict export controls.
Earlier this month, President Trump signed an executive order establishing a voluntary 30-day review process for top-tier private AI models. The order, a collaboration between the federal government and private sector entities, aims to streamline oversight while avoiding direct intervention in the industry. Trump had delayed signing the directive, stating he did not want to "get in the way" of U.S. leadership in AI or inadvertently benefit China. The executive order reflects a broader strategy to balance innovation with national security considerations.
Historical Context: Anthropic and the Pentagon Conflict
The current dispute with the Trump administration echoes a prior disagreement involving the Pentagon. Earlier this year, Anthropic clashed with the Department of Defense over its push to implement formal guardrails for the military’s use of its Claude models. These models, deployed in classified systems, were reportedly used to support autonomous weapons and mass surveillance. When negotiations stalled, Trump ordered federal agencies to halt their use of Anthropic’s technology, labeling it a "supply chain risk." Defense Secretary Pete Hegseth supported the move, arguing that the company’s approach could compromise national security.
Anthropic, long a proponent of government oversight in AI, criticized the administration’s actions as ideologically motivated. In a legal challenge, the company argued that the restrictions were "Orwellian" and aimed to "cripple" its operations. A federal judge sided with Anthropic, blocking the restrictions and calling them an overreach. The government is now appealing the decision, indicating ongoing efforts to assert control over AI deployment.
Broader Implications for AI Regulation
The resolution of this dispute highlights the evolving dynamics between AI developers and regulatory bodies. Anthropic’s claim that a narrow jailbreak should not justify recalling a widely used model raises questions about the standards for evaluating AI safety. If applied broadly, such a standard could halt new model launches, according to the company, which warns of a potential slowdown in innovation. Meanwhile, the Trump administration’s approach appears to prioritize industry collaboration over stringent mandates, as seen in the recent approval of Mythos 5 for select users.
The reinstatement of access to Fable 5 and Mythos 5 comes amid a broader conversation about the role of federal oversight in AI development. Critics argue that rapid deployment without thorough scrutiny could lead to unintended consequences, while supporters contend that regulatory intervention should not impede progress. The Trump administration’s willingness to engage with Anthropic suggests a shift toward a more flexible framework, though the underlying concerns about AI’s risks remain unresolved.
Stakeholders and the Path Forward
CBS News has contacted Anthropic and the Commerce Department for further clarification on the revised restrictions. The company’s social media announcement marks a key milestone, but the broader implications for AI regulation continue to unfold. With the federal government’s approval of Mythos 5 for limited testing, Anthropic is positioning itself to rebuild trust with key partners, while also advocating for a more balanced regulatory approach.
The Trump administration’s actions reflect a strategic effort to navigate the complexities of AI governance. By engaging with companies like Anthropic, it seeks to address security concerns without stifling innovation. The debate over export controls and model access underscores the challenges of regulating cutting-edge technology in a globalized world. As Anthropic prepares to restore access to its models, the question remains: will this resolution set a precedent for future AI policies, or will new disputes emerge as the technology continues to evolve?
The incident also highlights the interconnectedness of AI development and national strategy. With China’s advancements in AI technology a growing concern, the U.S. administration’s stance on regulating models like Fable 5 and Mythos 5 could influence international competition. Anthropic’s experience with the Trump administration serves as a case study in the delicate balance between safeguarding against misuse and fostering innovation. As the company moves forward, it will need to demonstrate that its models meet the government’s standards while maintaining their competitive edge in the global AI landscape.