# Advanced AI Society (AAI Society) # https://advancedaisociety.org ## About Advanced AI Society is the alliance for verifiable AI. We unite market leaders and deployers with insurers, policymakers, and civil society to standardize, deploy, and scale open verification, an AI security approach purpose-built for the agentic age. Verifiable AI is a latent market at the intersection of agentic AI, AI security, and AI data. The technology already exists: mature cryptography and decentralized trust infrastructure, now applied to the agentic stack. ## Purpose and mission Our purpose: A world where people can trust AI. Our mission: To make open verification the default for high-risk AI systems. How: We unite market leaders and deployers with insurers, policymakers, and civil society to standardize, deploy, and scale open verification, an AI security approach purpose-built for the agentic age. ## Key Definitions ### verifiable AI The part of AI Security concerned with evidence of what AI systems did, AI whose behavior can be verified against evidence, not taken on assertion. ### open verification In open verification, the root of trust is a mechanism anyone can verify, not a party anyone must believe. It is a category of approaches that is purpose-built for the agentic age, where deployment of non-deterministic agents requires mechanisms that remove the need to trust a single gatekeeper. Open verification carries the ethos of open source software into the agentic era, extending it from inspecting source code to verifying the actions agents take in the agentic stack. Open source won the enterprise on a security premise: you cannot secure what you cannot inspect. Open verification extends that premise: you cannot delegate what you cannot openly verify. Proof-of-Control is one way to reach open verification. ### Proof-of-Control Proof-of-Control is an industry-led open standard for openly verifying that an AI agent stayed within the controls it was given. Systems that conform to it produce tamper-evident evidence of the control-governed actions an agent took and whether each stayed within its controls, tied to a known operator, verifiable by anyone without trusting the operator or an auditor. The evidence spans the six domains of provenance, privacy, portability, authorization, identity, and security. It is binary and deterministic; an action either occurred and stayed in bounds, or it did not. It is the evidence that accountability, attribution, underwriting, and enforceable governance depend on. The standard does not confirm that an action's effects settled in the outside world, that a control was adequate, or that an action was wise, and it does not establish that the model itself is safe or deterministic. Coined and built by AAI Society; in development at the Open Verification Lab, hosted at LF Decentralized Trust. The v1.0 working draft is open for public comment through October 30, 2026. ### The Verifiability Gap The widening distance between what AI agents do and anyone's ability to openly verify that they stayed within the controls they were given. Coined by AAI Society. ### The Verifiability Tiers A four-level framework for verifying an agent's control adherence, ranked by trust. Tier 1 Assertion (the operator's word) and Tier 2 Attestation (a third party vouches) both ask you to trust a party. Past the binary threshold, Tier 3 Trust-minimized (anyone can verify) and Tier 4 Self-enforcing (it cannot run otherwise) ask you to take no party's word; the parties their soundness still rests on are disclosed rather than removed. ### How this relates to AI audit and AI assurance Open verification works alongside the assurance you already have. A Tier 2 audit rests on the operator's own record, and anchoring it in Tier 3 evidence gives the auditor something they did not have to take on anyone's word. When AI acts at machine velocity, periodic audits fall behind and an operator's log remains a claim, not evidence — which is why Proof-of-Control evidence is what accountability, attribution, underwriting, and enforceable governance depend on. ## The Six Domains of Verification Proof-of-Control covers control-governed actions across six domains: provenance, privacy, portability, authorization, identity, and security. ## The Two Pillars and Six Programs Advanced AI Society works through two pillars, three programs under each. The Open Verification Lab builds research, standard, and tools to make AI safety possible in deployment. Its programs: 01 Standard, 02 Research & tools, 03 Certification. Ecosystem Growth builds the category of verifiable AI. Its programs: 04 Policy, 05 Convene, 06 Insights. Three threads run through all six programs: Pilots, Insurance, and Field-building. ## The Seven Founding Beliefs 1. The Machine-to-Machine Era Demands Machine-Speed Trust. 2. Verification Cannot Belong to the Verified. 3. Open Verification Is the Architecture for Verifiable AI. 4. The Technology Is Here; Now We Must Make It Legible as Shared Public Infrastructure. 5. Building a Third-Party Open Verification Ecosystem Prevents Loss of Control. 6. Securing Machine Trust Is a Social Movement, Not Just a Technical Specification. 7. Pragmatic, Pluralistic Approaches Are the Way Forward. Full text: https://advancedaisociety.org/about ## Leadership - Tricia Wang, CEO & Co-founder - Michael Casey, President & Co-founder ## How to get involved For organizations, start-ups, and enterprises: - Membership (501(c)(6)): https://advancedaisociety.org/join#organizational For individuals: - Society Member (receive updates before the public, event discounts): https://advancedaisociety.org/join#member - Founding Contributor of Proof-of-Control (for security leaders, cryptographers, GRC professionals, and AI safety experts): https://advancedaisociety.org/join#founding-contributor - Research Fellow (join the Open Verification Lab to give your research project a home or take on one of our research areas): https://advancedaisociety.org/join#research-fellow Support: - Research Sponsor (companies, foundations, and individuals funding a named research question): https://advancedaisociety.org/join#sponsor ## Site map - / : Home — who we are, the two pillars and six programs, the case for verifiable AI, how to get involved - /why-now : Why now — the story of the Verifiability Gap and the open verification ecosystem - /proof-of-control : The Proof-of-Control standard, Verifiability Tiers, six domains, FAQ - /proof-of-control/insurance : The Proof-of-Control Insurance Working Group - /lab : The Open Verification Lab — standard, tools, research - /about : About Advanced AI Society — the seven Founding Beliefs, leadership, and advisors - /join : Membership, sponsorship, and founding contributor paths - /announcements : Announcements — news from Advanced AI Society; each post at /announcements/; RSS feed at /announcements/feed.xml - /briefing : Request a briefing - /contact : Contact form - /brand : Brand Guidelines for partners and press — logo downloads, usage rules, naming forms, and approved descriptions; printable PDF at /logos/advanced-ai-society-brand-guidelines.pdf - /antitrust : Antitrust Acknowledgement — the antitrust principles that apply at Advanced AI Society events and in the Proof-of-Control standard community ## Standard repository https://github.com/LFDT-ProofOfControl/ov-poc-standard