A common language forwhat AI is built from.
The AI Commons registry provides a shared account of the books, library archives, datasets, and models the field runs on. This includes provenance chains addressing where each came from, what is known about its origins, how it can be used, what it derives from, and what derives from it. We are building this with the institutions that have kept the web open, and publishing the design as we go, so the groups it describes can shape the process and results.
We track four core things about every entry.
The Registry records claims made about each work, which can include conflicting statements from different sources. In its first iteration it will not resolve any conflicts or differences in vocabulary. Over twenty access vocabularies exist for licensing and signalling preferences. Registry entries support claims made within the content of a work, by its authors, by its publishers and rightsholders, and by curators.
Authors, editors, curators; where it came from and who digitized or shared it.
Claims about how it can be used, stored in the work or in its environment.
Known derivations, upstream and downstream.
What we could not establish, stated explicitly.
To keep things simple, the registry is:
We are not asking anyone to adopt anything.
Position in the record cannot be bought. Nobody pays to rank higher, or to have their claims given primacy over others.
No membership fees are needed to have other data included.
This is a work in progress, and gaps will be marked as such.
We aren't trying to control the ecosystem, but to catalog and harmonize existing records, as a step towards better access and clarity for all creators, archivists, and reusers.
Compiled by us. Correctable by anyone.
We compile the record from what is already public. Nobody has to join it, sign anything, or opt in for it to be useful. An index is a claim about the world, not a membership list.
Every entry will carry a challenge link. Challenges will be recorded, attributed, and dated alongside the entry, whether or not we agree with them. Nothing gets quietly amended. This is the commitment that makes a compiled record legitimate rather than presumptuous, and it ships with the first version.
If you publish or steward an asset in the index, you will be able to claim its record: confirm what's right, amend what isn't, add what we missed. Claimed records will be marked as claimed.
A public registry is not ours to own.
Archives, libraries, universities, Creative Commons, Wikimedia and the open source communities have spent decades keeping this material available. They already describe it, each in their own way. The work is to agree on a shared way of saying it, and then to make sure the result can be mirrored anywhere rather than held in one place.
AI Commons is the place that this conversation happens.
A Global Data Pledge
The Global Data Pledge is a coalition of governments, memory institutions, publishers and rightsholder collectives committing to a consent-aware, provenance-rich commons. Its second ask (reciprocity) is recorded here, in this registry. That bridges the two: this commitment can be realized by contributing to a shared registry, or by maintaining local knowledge that a public registry can draw upon.
Read the Pledge →Progress
This initiative is currently underway with daily pledges, working groups drafting the technical standards and legal templates, and we plan to launch the live registry in Q1 2027. Long-term hosting, iterations on the framework, harmonization of terms, and how this is used by the ecosystem, will be ongoing work after that.