Anthropic commits $20m to political group backing state-level AI regulation
Anthropic will put $20 million behind Public First Action, a US political group supporting candidates who favor state authority to regulate AI. The move puts the company at odds with OpenAI's stance for less stringent, more centralized approaches.
Public First Action opposes federal efforts to preempt state AI laws, including a December executive order issued by Donald Trump. One of its endorsed candidates is Republican Marsha Blackburn, now running for governor in Tennessee, who pushed back against a congressional attempt to bar states from setting their own AI rules.
"The companies building AI have a responsibility to help ensure the technology serves the public good, not just their own interests," Anthropic said in a statement.
Why this matters for government professionals
This donation signals an intensifying policy split: state-level experimentation versus federal preemption. For agencies, that split will influence procurement terms, compliance timelines, and how fast new safeguards land in operations.
Public First Action was launched by two former members of Congress and serves as a counterweight to Leading the Future, a group generally opposed to strict AI rules. Leading the Future is backed by OpenAI president Greg Brockman and investor Marc Andreessen; Andreessen's firm, A16Z, is an OpenAI investor. The group has raised $125 million since August 2025, according to a spokesperson.
Expect more funding to shape AI policy in the midterms. Several states have already passed, or are considering, AI bills-meaning your agency could face divergent requirements across jurisdictions.
Key takeaways for agencies and public-sector teams
- Prepare for a patchwork: If federal preemption stalls, state rules will continue to drive requirements on transparency, testing, data governance, and model accountability.
- Update procurement: Add clear RFP clauses on model provenance, risk controls, audit logs, and incident reporting. Require vendors to disclose fine-tuning data sources and eval results where possible.
- Strengthen oversight: Stand up or expand an AI review board to assess high-risk use cases, bias testing, and red-teaming evidence before deployment.
- Budget for compliance: Allocate funds for monitoring tools, third-party assessments, and legal review to keep pace with new state laws.
- Coordinate early: Align with state CIO/CISO offices and attorneys general to avoid conflicts between federal guidance and state mandates.
What to watch next
- Campaign finance disclosures: Track who is funding which positions on preemption and safety standards.
- State legislative sessions: Timing of AI bills will dictate agency implementation windows and vendor obligations.
- Litigation risk: Expect challenges over federal preemption and constitutional questions around state AI oversight.
- Industry commitments: Watch for voluntary standards-or retreat from them-as funding lines harden.
Action checklist (near-term)
- Inventory AI systems, pilots, and vendor tools touching citizens, benefits, eligibility, law enforcement, and critical infrastructure.
- Map each system to current and proposed state requirements; flag gaps for remediation.
- Insert "regulatory change" clauses into contracts to shift costs or timelines when laws update.
- Launch training for program managers and procurement staff on AI risk, evals, and vendor due diligence.
Context and resources
- NCSL tracker: State activity on AI
- FEC overview: Types of political committees
- Practical AI courses by job role (procurement, policy, IT)
Bottom line
Anthropic's $20 million bet on state-led AI oversight raises the odds of divergent rules-and faster-moving requirements-for public agencies. Build capacity now, tighten procurement language, and keep a close eye on statehouses and court dockets.
Your membership also unlocks:
