California’s push to regulate frontier AI has gained an unusual ally from inside the industry. Anthropic has officially endorsed SB 53, a bill from state senator Scott Wiener that would set transparency requirements for the largest AI model developers.
The endorsement matters because major tech groups, including the Consumer Technology Association and Chamber of Progress, are lobbying against the bill. It also puts Anthropic on one side of a broader fight over whether AI safety should be handled by states, by the federal government, or by companies themselves.
What SB 53 Would Require
SB 53 is aimed at frontier AI model developers, including companies such as OpenAI, Anthropic, Google, and xAI. If passed, it would require those companies to create safety frameworks and release public safety and security reports before deploying powerful AI models.
The bill would also establish whistleblower protections for employees who raise safety concerns. That provision is important because the bill focuses on risks that may be hard for outsiders to see before a model is released or misused.
Senator Wiener’s proposal is not centered on everyday AI controversies such as deepfakes or sycophancy. Instead, it targets what the bill calls “catastrophic risks,” defined as the death of at least 50 people or more than a billion dollars in damages.
In practical terms, the bill is focused on the extreme end of AI risk. The source article describes concerns such as AI models providing expert-level assistance in creating biological weapons or being used in cyberattacks.
Why Anthropic’s Support Stands Out
Anthropic’s endorsement is notable because the company also says it would prefer federal AI rules. In its blog post, Anthropic said frontier AI safety is best handled at the federal level rather than through a patchwork of state regulations.
But Anthropic argued that waiting for Washington is not enough. The company said powerful AI advancements will not wait for federal consensus and described SB 53 as a solid path toward developing governance before a crisis forces a reaction.
Anthropic co-founder Jack Clark made a similar point in a post on X. “We have long said we would prefer a federal standard,” Clark said. “But in the absence of that this creates a solid blueprint for AI governance that cannot be ignored.”
That position gives SB 53 a rare form of support: approval from one of the companies that would likely be covered by the bill. It also separates Anthropic from other parts of the tech industry that are pressing California not to move forward.
The Opposition Is Still Strong
SB 53 faces resistance from both Silicon Valley and the Trump administration. Critics argue that state-level AI regulation could limit America’s innovation in competition with China.
Some opponents also say states should leave AI regulation to the federal government. Andreessen Horowitz’s head of AI policy, Matt Perault, and chief legal officer, Jai Ramaswamy, published a blog post arguing that many state AI bills risk violating the Constitution’s Commerce Clause.
The Commerce Clause argument, as described in the source article, is that states may not pass laws that go beyond their borders and impair interstate commerce. That objection goes directly to the heart of state AI rules because frontier AI companies operate across state lines.
OpenAI has also appeared in the debate. Its chief global affairs officer, Chris Lehane, sent a letter to Governor Newsom in August arguing that he should not approve AI regulation that would push startups out of California. The letter did not mention SB 53 by name.
OpenAI’s former head of policy research, Miles Brundage, criticized Lehane’s letter in a post on X, saying it was “filled with misleading garbage about SB 53 and AI policy generally.” The source article notes that SB 53 aims to regulate only the world’s largest AI companies, particularly those with gross revenue of more than $500 million.
How SB 53 Differs From Earlier Efforts
California’s Senate has approved a prior version of SB 53, but a final vote is still needed before the bill can advance to Governor Gavin Newsom’s desk. Newsom has not taken a public position on SB 53 so far, though he vetoed Senator Weiner’s last AI safety bill, SB 1047.
Policy experts quoted in the source article describe SB 53 as more modest than earlier AI safety bills. Dean Ball, a senior fellow at the Foundation for American Innovation and former White House AI policy adviser, said in an August blog post that SB 53 has a good chance of becoming law.
Ball had criticized SB 1047, but said SB 53’s drafters showed “respect for technical reality” and a “measure of legislative restraint.” That assessment suggests the bill has been shaped to answer some earlier objections while still putting legal obligations around safety practices.
Senator Wiener has said SB 53 was heavily influenced by an expert policy panel convened by Governor Newsom. That panel was co-led by leading Stanford researcher and co-founder of World Labs, Fei-Fei Li, and was created to advise California on how to regulate AI.
From Voluntary Promises To Legal Duties
Many leading AI labs already maintain internal safety policies and publish safety reports. The source article names OpenAI, Google DeepMind, and Anthropic as companies that regularly publish reports for their models.
The issue is that those commitments are currently self-imposed. Companies are not bound by anyone but themselves, and the source article notes that they sometimes fall behind their own safety commitments.
SB 53 would shift some of those practices from voluntary company policy into state law. It would also attach financial repercussions if an AI lab fails to comply.
California lawmakers amended SB 53 earlier in September to remove a section that would have required AI model developers to be audited by third parties. Tech companies have fought those kinds of third-party audits in other AI policy battles, arguing that they are overly burdensome.
The result is a bill that still targets major AI risks, but does so through public reporting, safety frameworks, whistleblower protections, and legal accountability. Anthropic’s endorsement does not end the fight over SB 53, but it changes the politics of the debate by showing that at least one major frontier AI developer is willing to support state-level rules while federal action remains uncertain.