AI Safety Institute METR Secures $71 Million in Funding Commitments to Expand Autonomous Risk Audits

Model Evaluation and Threat Research, the independent artificial intelligence safety nonprofit known as METR, announced on Friday that it has raised approximately $71 million in new funding commitments over the past six months. Operating from its primary research hub in downtown Berkeley, the organization will deploy the philanthropic capital to significantly expand its technical engineering staff and launch ambitious new evaluation programs across the Bay Area. The substantial funding infusion arrives as federal and state regulators increasingly rely on third-party auditors to verify the safety claims of major commercial AI developers.

Leadership officials confirmed that the $71 million in commitments will directly fund five core technical pillars over the coming multi-year cycle: studying autonomous agent capabilities, tracking recursive self-improvement in frontier models, evaluating automated monitoring systems, conducting pre-deployment risk assessments, and investigating real-world AI security incidents. As autonomous coding and scientific reasoning systems grow more complex, METR researchers are building standardized testing environments to measure whether models can autonomously replicate, evade human oversight, or accelerate cyber offense.

To preserve absolute institutional objectivity during high-stakes audits, METR emphasized its strict conflict-of-interest governance policies regarding commercial industry ties. The nonprofit has not accepted any financial funding from frontier AI companies, nor does it accept charitable donations made by or at the direction of staff employed at frontier AI laboratories. While METR receives free compute tokens directly from model developers to execute capability evaluations, scientific research, and harness engineering, its operational budget remains entirely insulated from commercial laboratory influence.

Technology policy analysts across Berkeley, Oakland, Walnut Creek, and San Francisco noted that independent verification institutes play a vital civic role in the Bay Area's rapidly expanding artificial intelligence ecosystem. University faculty and municipal advisory boards observed that having a well-capitalized, financially independent auditing body in Alameda County strengthens public accountability as San Francisco and Silicon Valley labs release increasingly autonomous software. The organization is currently recruiting senior distributed systems engineers, security auditors, and machine learning scientists across the East Bay.

Researchers, philanthropic partners, and software engineers interested in METR's conflict-of-interest disclosures, governance board structure, and technical job openings can review full documentation online at metr.org or call the Shattuck Avenue office in Berkeley at (510) 647-3900. Public methodology papers detailing recursive self-improvement benchmarks and control evaluations are published regularly on the institute's digital portal. Questions regarding independent audit partnerships or academic collaborations may be directed to the external affairs committee.