Anthropic Responds to California Governor Newsom’s AI Working Group Draft Report

TL;DR

Anthropic praised Governor Newsom’s AI Working Group draft report for emphasizing objective standards and transparency, and highlighted that the lab already follows many of the report’s recommendations through its Responsible Scaling Policy, public safety disclosures, and third‑party testing, while advocating for light‑touch government measures to make such practices universal.

Alignment with the Working Group’s Focus

Anthropic agrees that objective, evidence‑based standards are essential for frontier AI governance. The lab specifically supports the report’s emphasis on transparency as a low‑cost, high‑impact tool for:

  • Expanding the evidence base around new AI technologies.
  • Building consumer trust.
  • Encouraging positive‑sum competition among AI companies.

Anthropic welcomed the Working Group’s call for labs to disclose model theft protections and national‑security risk testing.

Existing Anthropic Practices that Mirror Report Recommendations

  • Responsible Scaling Policy – publicly details how Anthropic assesses misuse and autonomy risks, and defines thresholds that trigger heightened safety and security measures.
  • Safety and security testing disclosures – results are shared with each major model release.
  • Third‑party testing – external audits complement internal evaluations, as described in Anthropic’s third‑party testing announcement.

These practices align with the report’s recommendation that frontier labs adopt transparent safety and security protocols.

Call for Light‑Touch Government Intervention

Anthropic notes that, currently, frontier AI firms are not required to:

  1. Publish a safety and security policy.
  2. Describe that policy publicly.
  3. Document the tests they conduct.

The lab argues that modest, light‑touch regulation could close this gap without stifling innovation. In a recent policy submission to the White House, Anthropic warned that powerful AI systems could appear as early as the end of 2026, underscoring the urgency of establishing a transparent policy regime.

Economic Impact Focus and Future Contributions

The Working Group highlighted the need for deeper research on AI’s economic effects. Anthropic is responding by developing an Economic Index to track AI‑driven economic changes. The lab plans to continue providing feedback to the Working Group to help finalize the report.

Conclusion

Anthropic commends Governor Newsom’s initiative to launch the AI Working Group and commits to collaborating on California’s approach to frontier model safety, while urging broader industry adoption of transparent safety and security practices through light‑touch regulation.

Sources

Related

  • Dispatch
    Anthropic Endorsement of California SB 53

    Anthropic has endorsed California Senate Bill 53 (SB 53), a regulatory framework for powerful AI systems that prioritizes transparency and disclosure over prescriptive technical mandates.

  • Dispatch
    Anthropic Third-Party Testing AI Policy Proposal

    Anthropic proposes a third-party testing regime for frontier AI systems to validate safety, prevent national security risks, and avoid regulatory capture through a diverse ecosystem of auditors.

  • Dispatch
    Anthropic Response to America's AI Action Plan

    Anthropic supports the White House's AI Action Plan for its focus on infrastructure and federal adoption but urges stronger national transparency standards and stricter export controls on H20 chips to China.

  • Dispatch
    Anthropic Proposal for Increased NIST Funding for AI Measurement

    Anthropic proposes ambitiously funding the National Institute of Standards and Technology (NIST) to develop standardized AI measurement tools and safety thresholds, which they argue is a prerequisite for effective AI regulation.

  • Dispatch
    Anthropic Proposed AI Development Transparency Framework

    Anthropic has proposed a targeted transparency framework for the largest AI developers to standardize safety disclosures and ensure public accountability without impeding innovation.