California lawmakers call for emergency laws and criminal penalties as AI leaders warn of runaway systems

# What happened

# Why lawmakers are acting now

Representative Ro Khanna (D-Fremont) argued companies should face liability or criminal sanctions if their systems do illegal things. He and other Democrats want immediate legislation to create legal consequences for creators of "rogue" AI.

# Where elected officials stand

• Democratic lawmakers including Khanna and Rep. Ted Lieu urged an emergency session of Congress to impose guardrails after several AI executives acknowledged safety risks. Lieu criticized the prior federal approach and urged swift federal action.
• California Gov. Gavin Newsom has pushed for national legislation and recently signed bills to create pathways for outside audits of top AI companies.
• House Speaker Mike Johnson and former President Trump resisted immediate action. Johnson warned that rushing regulation could "smother American innovation" and harm competitiveness with China. Trump publicly downplayed concerns about the pace of AI progress.

# What tech leaders said

Dario Amodei said the industry should "slow the pace," arguing the potential for self-improving models is serious enough to justify restraints. OpenAI, Anthropic, and executives including Sam Altman and Elon Musk publicly supported calls for slower development or stronger guardrails in recent days. A former researcher who left Anthropic also posted that he resigned after concluding companies were racing toward self-improving superintelligence.

# Concrete policy ideas being proposed

Lawmakers and advocates have floated measures that include:

• unlawfully.
• Emergency legislation to require monitoring and safety controls during model training and evaluation.
• Mechanisms for outside audits of frontier models, a step California has already started through state legislation.

OpenAI and others have urged strengthening existing state law frameworks to require monitoring of models under training to detect incidents that could bypass security controls.

# Public reaction and political context

Public concern about AI is increasing. The piece cites a Gallup poll finding that 7 in 10 Americans oppose local construction of data centers used to power AI. The debate over regulation is entangled with broader political concerns about U.S. competitiveness with China and past federal approaches to tech oversight. California is moving ahead with state-level tools while pushing for a federal response.

# What to watch next

• Whether Congress reconvenes for an emergency session or advances federal guardrails.
• Any California bills that expand auditing or liability requirements beyond current measures.

Overall, the story captures an intensifying clash: lawmakers demanding immediate legal levers to contain risks described by AI executives, while some federal leaders warn regulation could undercut U.S. innovation in a global race.