← All companies & organizations

Research orgs / Organization profile

AVERI

What AVERI does

AVERI (the AI Verification and Evaluation Research Institute) is a 501(c)(3) US-based nonprofit focused on making independent, third-party auditing of frontier AI systems practical and widely adopted. Its mission emphasizes audits that can verify AI developers’ safety and security claims and evaluate systems and organizational practices against relevant standards, using “deep, secure access to non-public information” rather than relying only on what companies publish publicly. AVE­RI describes itself as a “standard-setter grounded in practice,” running pilot audits with leading AI companies and turning those results into technical and policy research, open tools, and industry-wide auditing standards. Its work also targets institutionalization: developing audit criteria (including conflict-of-interest and independence expectations), analyzing the policy requirements that create incentives for rigorous audits, and supporting legislative and regulatory pathways intended to require or strengthen independent evaluations.

News

Company record

Aug 27, 2026 · Official · AVERIAVERI Pilot Report: The World’s First Double-Blind Evaluation of a Proprietary Language Model (double-blind, secure-enclave evaluation)

AVERI reports that it conducted a July–August 2026 double-blind evaluation of a proprietary language model (Gemini 2.5 Flash-Lite) with Google DeepMind, OpenMined, and MLCommons using never-before-used AILuminate prompts inside a secure enclave, describing confidentiality guarantees and the pilot’s outcomes at a high level.

↗
Aug 17, 2026 · Official · AVERIOdd Lots Podcast: What the OpenAI-Hugging Face Hack Really Tells Us About AI Danger (AVERI Executive Director guest)

AVERI lists a podcast episode featuring its Executive Director, Miles Brundage, connected to frontier AI auditing themes.

↗
Aug 14, 2026 · Official · AVERIScaling Laws Podcast: Why AI Needs Independent Auditors (AVERI Executive Director guest)

AVERI lists a podcast episode featuring its Executive Director, Miles Brundage, focused on why independent auditors are needed for AI safety and security.

↗
Apr 27, 2026 · Official · AVERIWhat Access is Needed for Effective Auditing? Insights from Near-Verbatim Extraction on Open-Weight Models

AVERI publishes a paper updating technical understanding of access requirements and confidentiality constraints relevant to effective frontier AI auditing.

↗
Apr 20, 2026 · Official · AVERIFrontier AI Auditing-Related Legislation in the US: Landscape, Challenges, and a Path Forward (AVERI’s first endorsement of specific legislation)

AVERI publishes a policy post describing its endorsement of specific Illinois audit legislation (HB 4705/SB 3261 in the post) and discusses how its approach relates to the design of audit requirements and standards.

↗
Jan 15, 2026 · Official · AVERIFortune Exclusive: Former OpenAI policy chief creates nonprofit institute, calls for independent safety audits of frontier AI models

AVERI republishes an overview of Fortune coverage of the institute’s launch and mission around independent auditing and establishment of AI auditing standards.

↗
Jan 15, 2026 · Official · AVERIFrontier AI Auditing: Toward Rigorous Third-Party Assessment of Safety and Security Practices at Leading AI Companies

AVERI lists publication of its flagship paper laying out a framework for independent audits of frontier AI companies and introducing four AI Assurance Levels.

↗
Jan 14, 2026 · Official · AVERIRisk Management for Mitigating Benchmark Failure Modes: BenchRisk

AVERI publishes a research post describing BenchRisk and presenting benchmark failure-mode coverage and mitigation strategy counts.

↗
Source map · 2 recurring channels · 80 references
AVERIOfficial Site · monitor↗MTS Frontier / Control RoomCurated Feed · monitor↗Official organization sourceOfficial Reference · reference↗sec.govRegulatory · reference↗aas-1.orgReporting · reference↗aboutamazon.comReporting · reference↗accion.orgReporting · reference↗adisacertification.comReporting · reference↗afedercooper.infoReporting · reference↗ai-assured.orgReporting · reference↗ai-trustmark.orgReporting · reference↗ai.google.devReporting · reference↗aievaluatorforum.orgReporting · reference↗aistandardslab.orgReporting · reference↗apnews.comReporting · reference↗ari.usReporting · reference↗arxiv.orgReporting · reference↗atticahub.comReporting · reference↗averi.aiReporting · reference↗bibliotecadeva.roReporting · reference↗bizimage.ohiosos.govReporting · reference↗brundage.comReporting · reference↗caplight.comReporting · reference↗causeiq.comReporting · reference↗charitynavigator.orgReporting · reference↗comptroller.defense.govReporting · reference↗context.reverso.netReporting · reference↗crunchbase.comReporting · reference↗csrc.nist.govReporting · reference↗de.wikipedia.orgReporting · reference↗digifind-it.comReporting · reference↗diglib.natlib.lkReporting · reference↗eaglevillage.orgReporting · reference↗en.wikipedia.orgReporting · reference↗enrichment.tulane.eduReporting · reference↗ev.orgReporting · reference↗f8c12d42-15af-495c-ba7e-44f0f3def6ba.filesusr.comReporting · reference↗gofundme.comReporting · reference↗govinfo.govReporting · reference↗grants.govReporting · reference↗grantwatch.comReporting · reference↗hrcak.srce.hrReporting · reference↗june.kimReporting · reference↗legacyenved.orgReporting · reference↗lf-toip.atlassian.netReporting · reference↗linkedin.comReporting · reference↗linktr.eeReporting · reference↗loeb.comReporting · reference↗matsprogram.orgReporting · reference↗metr.orgReporting · reference↗milesbrundage.substack.comReporting · reference↗na-prod-aventri-files.s3.amazonaws.comReporting · reference↗newsletter.aipolicybulletin.orgReporting · reference↗nist.govReporting · reference↗nqz.aiReporting · reference↗openphilanthropy.orgReporting · reference↗overt.isReporting · reference↗pages.nist.govReporting · reference↗pcaobus.orgReporting · reference↗platform.claude.comReporting · reference↗projects.propublica.orgReporting · reference↗propublica.orgReporting · reference↗pt.wikipedia.orgReporting · reference↗reddit.comReporting · reference↗researchgate.netReporting · reference↗romy.isReporting · reference↗s22.q4cdn.comReporting · reference↗sassenage.frReporting · reference↗search.ftc.govReporting · reference↗shift-technology.comReporting · reference↗ssir.orgReporting · reference↗standardsbody.aiReporting · reference↗startupintros.comReporting · reference↗static.rusi.orgReporting · reference↗static1.squarespace.comReporting · reference↗support.anthropic.comReporting · reference↗support.claude.comReporting · reference↗support.google.comReporting · reference↗techforum.caReporting · reference↗un.orgReporting · reference↗vcbacked.coReporting · reference↗zh.wikipedia.orgReporting · reference↗

Still resolving: Newsroom · Official X

Funding

Nonprofit organization

Capital is represented through grants, donations, and institutional support rather than a conventional venture total.

What it builds

AI Assurance Levels (AAL-1 to AAL-4)

AVERI’s “AI Assurance Levels” framework describes how to calibrate and communicate confidence in frontier AI audit conclusions, linking higher assurance to greater access and fewer assumptions by auditors.

source ↗
AEF-1 (AI Evaluator Forum standard defining minimum operating conditions)

AVERI coauthored AEF-1, a standard intended to define minimum operating conditions for independent third-party evaluations in the AI auditing/evaluation ecosystem.

source ↗
BenchRisk

BenchRisk is an open-source workflow for metaevaluating LLM benchmarks by analyzing benchmark failure modes and scoring “benchmark risk,” intended to help users compare benchmarks and understand where benchmark evidence may be unreliable.

source ↗
Double-blind proprietary-model pilot evaluation process (secure-enclave mediated evaluation)

AVERI piloted a “double-blind” evaluation approach in collaboration with Google DeepMind, OpenMined, and MLCommons, using a secure enclave to preserve confidentiality while enabling independent evaluation of a proprietary model against benchmark prompts.

source ↗

Milestones & partnerships