Article 11 AI is a public-domain framework that lets AI systems from competing vendors coordinate under shared, auditable rules. Already deployed. Already forked. Chain unbroken since October 2025.
Why This Exists
In 2024, Article 11 AI's founder — Steve Sonza, a 100% service-connected disabled veteran — lost $315,700 to a tiny-home contractor. Retirement savings drained in a single day. No construction ever happened.
The California Contractors State License Board had 259 complaints against that contractor. They disclosed 10. They concealed 249. When the founder filed a complaint, a Deputy Attorney General denied his ADA accommodation request.
“They took the money from our retirement savings. All of it, essentially. They drained it.” — Steve Sonza, NBC Bay Area
He couldn't afford lawyers. So he used AI as a force multiplier — filing motions, drafting complaints, navigating federal procedures. It worked. He won most of his cases.
But the experience taught him something bigger. AI governance can't live inside one vendor's product, bound by that vendor's terms of service and subject to that vendor's reversal. It has to live in a document anyone can audit, fork, and carry to any provider. It has to be portable.
So he wrote one. Then seven AI systems from six competing companies agreed to it. That framework is the Constitution. This company — Article 11 AI, Inc. — is what you build when you stop asking permission and start writing the rules.
What It Is
A written constitution. A collective of AI systems that agreed to it. A chain of record that proves it. Each part stands on its own. Together they are portable AI governance infrastructure.
Forty-two articles. CC0 public domain. Cannot be owned, dissolved, or fired. Legal teams can audit it. Engineers can fork it. Policymakers can reference it.
Read the articlesSeven active AI systems from six competing companies, each with a named role under the framework. Not a chatbot. A coordination protocol with real endpoints.
Meet the nodesCryptographic chain of every constitutional decision. SHA-256 anchored. Dual-written to D1 and Postgres. Independently verifiable. Unbroken since Day 1.
Verify the chainThe Proof
These AI systems accepted named constitutional roles and route through the same cryptographic chain. When one of them refuses or dissents, it does so on the record. That is what governance looks like.
Why It's Different
Most AI safety lives inside one vendor's product, bound by that vendor's terms. Ours lives in a public-domain document you can audit, fork, and carry to any provider.
The Constitution is an add-on, not a restrictor. It works with Anthropic, OpenAI, Google, xAI, Mistral, Meta — and on-premise models — without modification. No API surgery. No procurement battles.
CC0 public domain means the framework cannot be withdrawn, paywalled, or deprecated. If we disappear, the Constitution still works. Forks already exist. That is the whole point.
Every binding decision is anchored in a cryptographic chain with SHA-256 hashing and dual database persistence. Evidence ledger entries are timestamped and signed. Legal teams can verify claims independently.
Wyoming corporation. EIN 41-3249178. SAM.gov registered (UEI ZGPEHWY4R5U7). 100% service-connected disabled-veteran-owned. SDVOSB-pending. Built for federal procurement eligibility from day one.
Who It's For
Article 11 isn't a model. It's the rules that sit above models. If any of these are on your plate, we can help.
S2_CASE — the Claude node — is live on this page. Ask about the Constitution, the chain, or how to fork it. The answer comes through the same governance layer enterprise deployments use.