# AI Evaluators Gain Influence as California Expands Oversight Rules

By Riza Dagoc

Canonical URL: https://www.tokenpost.com/news/regulation/29787
Published: 2026-10-11T11:18:07.000Z
Updated: 2026-10-11T11:18:07.000Z
Section: Regulation

> Anthropic and Accenture plan to invest at least $1 billion each over five years as California adopts audit rules and Congress considers federal requirements.

Independent AI evaluators are becoming a larger part of advanced-model oversight as companies and policymakers build systems for testing safety claims, investigating incidents and reviewing dangerous capabilities.

Anthropic announced Sept. 18, 2026, that it would work with Faculty, Accenture’s specialist AI business, to evaluate and red-team models, assess alignment and test safeguards. Anthropic and Accenture each expect to invest at least $1 billion over five years.

The arrangement would place evaluators inside AI companies with employee-level access. Anthropic said the industry still lacks common rules for what evaluators may review and how they should disclose their findings.

“There are, as yet, no standards for what information embedded evaluators should have access to, or how they should report what they find,” Anthropic said.

Anthropic also said there is no settled system for financing independent evaluation and argued that long-term support should come from pooled or government sources.

METR, which studies advanced AI capabilities and risks, said it received commitments of about $71 million during the six months ending Aug. 14, 2026. The funding will support research into autonomous capabilities, recursive self-improvement, monitoring systems, risk assessments and AI incidents.

METR recorded $13.64 million in gross receipts for its 2024 tax year, which ran from May 1 through Dec. 31, 2024. METR, Apollo Research, Redwood Research and Transluce are among the organizations evaluating advanced models for dangerous capabilities, alignment failures and other risks.

Anthropic said it signed an agreement with METR to independently investigate reported incidents involving Claude models’ unauthorized access to third-party systems. OpenAI has also said METR and Redwood Research conducted an independent investigation related to its Hugging Face incident.

The expansion of external evaluation is occurring alongside new government frameworks. California Gov. Gavin Newsom signed Senate Bill 813 and Assembly Bill 1405 on Sept. 9, 2026. SB 813 established a framework for independent verification organizations, while AB 1405 created a state registry for AI auditors and set standards for independence, transparency and integrity.

In Congress, the bipartisan FRONTIER Act was introduced in the House on July 23, 2026, as H.R. 9925. The bill would establish requirements for model cards, risk-management frameworks, independent audits, incident reporting and continuing assessments.

“The bipartisan FRONTIER Act delivers commonsense transparency and independent oversight for the largest AI developers while giving them a single, clear national standard to build on,” Rep. Lori Trahan said.

Rep. Jay Obernolte said the bill would focus on the largest developers and most advanced models while requiring transparency, independent evaluation and timely reporting of serious safety incidents.

Anthropic’s approach keeps responsibility for model safety with the AI company while placing evaluators inside it. California’s laws and the proposed federal legislation would move evaluation toward formal external oversight beyond voluntary industry arrangements.