Skip to content
AI Commons · Est. 2016

The knowledge that makes AI possible
came from all of us.
Every person has the right to benefit.

AI Commons is building the shared coordination and trust layer for consent-aware, provenance-rich AI. The missing infrastructure that makes trustworthy AI possible at scale.

Read the manifesto Our projects →
45+
Partners & collaborators
10
Years of foundational work
4
Active projects
What we're building

Four projects, in active development.

AI Commons is building four interlocking pieces of public infrastructure for consent-aware, provenance-rich AI. AMPL is the license layer. The Registry is the assertion substrate. The Commons Dataset is what the federation produces when the two combine. The Global Data Pledge is the front door through which institutions contribute.

01
AMPL
AMPL is an OSI-compatible base license for AI models and datasets, with four optional modules that handle the obligations that software licenses don't cover. Currently at v0.9 public draft, in legal review with the inner core.
02
Registry
The Provenance Registry is a machine-readable record of the status, provenance, and consent terms attached to works in the open web. Two layers are operational: a Publisher Registry where institutions declare their position on AI training, and an Assertion Registry where custodial attestations are recorded.
03
Commons Dataset
The Commons Dataset brings together existing open training corpora into a single AMPL-aligned, registry-anchored federation. AIC contributes the licensing infrastructure, the provenance coverage, and the coordination between corpus teams.
04
Global Data Pledge
Institutions pledge their data and preferences into the commons. Each pledge becomes a citable public commitment in AMPL vocabulary, anchored in the Registry. Active campaign targeting 20 institutional pledges by 7 July 2026 at the AI for Good Summit.
See all projects →
Who this is for

The commons belongs to everyone.

Funders

Put resources into shared infrastructure

Not just frontier races. The commons returns value to every actor in the ecosystem.

Builders

Publish your lineage. Respect permissions.

Clear training data means lower legal exposure and faster institutional adoption.

Governments

Require provenance and evidence

AI Commons builds the operational layer that makes requirements enforceable.

Creators

You are the source

You have the right to know how your work was used, and to challenge systems that affect your life.

Our network

45+ partners across
six continents.

"AI did not come from nowhere. It came from us. All of us. Which means we are not outsiders asking for access. We are the source."
AI Commons Manifesto · March 2026 · Read the full manifesto →
From the blog

Writing & updates.

Coming soon
Why the commons needs a coordination layer
The pieces of trustworthy AI infrastructure exist. The problem is they don't talk to each other.
Coming soon
What AMPL is and why it matters
A plain-language explanation of the AI Commons licensing framework and what problem it solves.
Coming soon
AI Commons 2.0: what we articulated at Berkman Klein
In December 2025, we spent two days with MIT Media Lab, Creative Commons, Wikimedia, and others. Here's what emerged.