Anthropic Backs California AI Safety Bill Amid Tech Industry Resistance

San Francisco, CA — AI company Anthropic has publicly endorsed California's SB53 bill, proposed by Governor Scott Wiener, marking a significant step toward regulating powerful AI models. The bill introduces unprecedented transparency requirements for major AI developers but faces strong opposition from Silicon Valley and federal stakeholders.

The SB53 Bill: A Landmark Proposal

The legislation targets developers like Anthropic, OpenAI, Google, and xAI, requiring them to:

• Develop safety frameworks before deploying advanced AI models.
• Publish public safety and security reports.
• Protect employees who report safety concerns.

The bill focuses on mitigating "catastrophic risks," defined as events causing at least 50 deaths or $1 billion in losses. It aims to prevent extreme scenarios, such as AI-assisted biological weapon development or cyberattacks, though it does not address immediate issues like deepfakes.

Anthropic’s Stance: Urgency Over Federal Delay

In a blog post, Anthropic acknowledged that while federal oversight would be ideal, "the development of powerful AI technology will not wait for consensus in Washington." The company praised SB53 as a "reasonable path" toward establishing governance standards.

Jack Clark, Anthropic’s co-founder, emphasized the bill’s importance: "While we hope for federal standards, SB53 provides an indispensable blueprint for AI governance."

Opposition from Tech and Federal Leaders

The bill has drawn criticism from:

• Silicon Valley investors, including Andreessen Horowitz and Y Combinator, who argue state-level regulation could hinder U.S. competitiveness with China.
• The Trump administration, which opposes what it views as overreach into innovation. Opponents contend that AI safety should remain a federal matter.

Legislative Progress and Challenges

The California Senate has passed SB53’s preliminary version, but it awaits a final vote before reaching Governor Gavin Newsom’s desk. Newsom previously vetoed a similar bill (SB1047) but has yet to take a position on SB53.

Policy experts note SB53 is more moderate than earlier proposals, reflecting legislative restraint. Its fate hinges on balancing safety concerns with industry pushback.

Key Points:

• 📜 Anthropic supports SB53, calling it critical for AI governance.
• ⚖️ The bill mandates safety frameworks, public reports, and whistleblower protections.
• 🔍 Focuses on preventing extreme risks (e.g., bioweapons), not immediate harms like deepfakes.
• 🏛️ Faces opposition from tech investors and federal leaders favoring innovation over regulation.