US Regulators to Leverage Existing Laws to Police AI Risk

Overview

The US government is considering using existing liability laws to regulate artificial intelligence, a move that has sparked debate among experts.

Administration's Approach

President Donald Trump's administration favors enforcing existing liability laws to manage the technology's risks.

Backing from Prominent Figures

The stance has backing from prominent names, including Treasury Secretary Scott Bessent and David Sacks, the former White House AI czar.

Non-Binding Accord

The White House meeting on September 29-30, 2026, endorsed a non-binding accord among major AI companies, including Anthropic, OpenAI, Nvidia, Meta and Google.

Independent Audits

The accord focuses on two things: robust internal processes at the companies, and independent audits.

Expert Disagreement

Experts cannot agree on what they should look like, with some arguing that self-regulation is insufficient given the technology's critical nature.

RAND Report

A RAND report from March 2026 found little optimism that strict regulation is achievable, with experts rating incentives for risk disclosure as the most promising path.

State-Level Regulation

Governor Gavin Newsom issued an executive order launching a review of AI safety plans on September 18, 2026.

Implications for the Industry and Investors

The administration's approach offers breathing room for companies building frontier AI, but relying on liability law means the boundaries get defined after harm occurs, often through lawsuits.

Watching the Developments

Experts are watching the independent audits promised under the accord, California's review of AI safety plans, and the potential for dissent to spread among executives.