Anthropic Pushes States for Tougher AI Regulations, Sparking Debate Over Motives

Anthropic, the AI startup valued at nearly $1 trillion, is urging state lawmakers across the United States to adopt stricter regulations on powerful AI systems, arguing that existing transparency laws are already insufficient to address the rapidly evolving technology.

The company, which helped secure transparency requirements in California and New York last year, now believes those measures fall short. Cesar Fernandez, Anthropic's head of US state and local government relations, told WIRED that transparency and self-reporting are "no longer sufficient safety measures for the most powerful AI systems."

From Transparency to Third-Party Audits

Anthropic's evolving policy stance marks a shift from the transparency-focused bills it championed last year. Fernandez said the policy responses need to match the pace at which AI capabilities are advancing.

Beyond the self-reporting laws in California and New York, the company has backed an Illinois measure that would require AI labs to have their safety processes evaluated by third-party auditors. More recently, Anthropic endorsed a Massachusetts policy that would mandate similar third-party auditing and empower the state's attorney general to seek injunctive relief against non-compliant companies.

Fernandez, who joined Anthropic earlier this year after stints at FanDuel and Uber, brings experience navigating state-level policy battles. His expertise comes at a time when Congress has failed to pass comprehensive AI regulation, leaving states to take the lead.

Anthropic's leaders have long maintained that the company must build a massive business around advanced AI to fulfill its founding mission of ensuring a safe transition through transformative AI. The regulations the company supports are designed to mitigate catastrophic risks, including the possibility that advanced models could contribute to financial disasters or mass deaths.

Accusations of Regulatory Capture

While AI safety groups and labor unions have praised Anthropic's pro-regulation stance, some Silicon Valley figures view the company's efforts with suspicion. David Sacks, the former White House AI czar and current technology adviser to President Donald Trump, accused Anthropic of running a "sophisticated regulatory capture strategy based on fear-mongering" in a social media post last year. Sacks claimed the company is "principally responsible for the state regulatory frenzy that is damaging the startup ecosystem."

Fernandez rejected the accusation, pointing out that Anthropic has only supported bills targeting "large AI model developers"—generally defined as companies that have spent hundreds of millions of dollars on AI development and generate more than $500 million in annual revenue. He argued it is difficult to imagine a startup meeting that threshold.

However, in an industry where developing competitive frontier AI models requires billions in capital, a handful of emerging companies could eventually cross those thresholds. Safe Superintelligence, Thinking Machines Lab, and Mistral have each raised billions from investors, though their revenue remains well below that of Anthropic or OpenAI. While these companies are not typical startups, they are potential competitors.

Fernandez countered that any company large enough to develop a powerful AI model should face the same regulatory requirements, as the underlying risks remain identical regardless of the developer. He said part of Anthropic's goal is to "inspire a race to the top developing the most safe and secure AI systems."

Where Anthropic Draws the Line

There are limits to what Anthropic believes states should do. In a policy document published last month, the company recommended that governments have a mechanism to block companies from deploying AI models deemed unsafe—but it argues that authority should rest with the federal government, not state lawmakers.

The suggestion carries some irony. The Trump administration recently directed Anthropic to suspend access to its two most powerful AI models, known as Mythos and Fable 5, for foreign nationals. The export control directive forced the company to take those models offline for all users. In a blog post, Anthropic argued that blocking AI model deployments should only occur through a fair and transparent evaluation process.

Fernandez described the question of which level of government should hold blocking authority as an "evolving conversation."

On the federal front, Anthropic has also been active. Last month, the company sent a letter to the US government accusing Chinese tech giant Alibaba of conducting a "distillation attack"—systematically extracting information from Anthropic's models through prompting to develop its own AI tools. Some AI researchers dismissed the claims as another form of regulatory capture, arguing that Anthropic's true objective is to persuade the US government to ban open-weight Chinese models. Such a ban could push thousands of American businesses that rely on those models toward Anthropic's enterprise products. Anthropic CEO Dario Amodei has previously warned Congress about the dangers of open source AI.

At the state level, Fernandez said Anthropic has not specifically targeted open source models in the legislation it has endorsed. He emphasized that the focus should be on model capabilities rather than how models are constructed.

A Narrow Focus on Catastrophic Risk

Anthropic's state-level advocacy has concentrated heavily on catastrophic and existential AI risks. The company's recent model releases have drawn national attention to the cybersecurity capabilities of advanced AI, and it has spent years warning lawmakers about potential disasters.

However, the concerns most frequently raised by American voters—such as job displacement, the impact of data centers on local communities, and the effects of chatbots on children—have not generated the same coordinated legislative push from frontier AI labs. Anthropic and several rivals have pledged that ordinary taxpayers will not bear the cost of data centers, and Anthropic has published proposals for addressing future AI-driven job losses. Still, the industry has devoted far less political effort to state laws tackling those issues.

Fernandez said Anthropic is "eager to engage with lawmakers" on topics beyond catastrophic safety risks and is already "having those conversations throughout the states." Those discussions, however, have not yet produced the kind of organized state-level campaign the company has mounted around existential threats posed by frontier models.

Whether Anthropic's regulatory push is driven by genuine safety concerns or business strategy, the company is undeniably shaping the future of AI policy in the United States. As states continue to fill the void left by congressional inaction, the debate over who should regulate AI—and how strictly—will only intensify. What do you think about Anthropic's approach to AI regulation? Share this article and join the conversation.

Source: Wired