A commitment to make accessible data available,and to make reciprocity real.
We, the undersigned, commit to making the publicly accessible data our institutions hold available for AI training, and to doing so under terms that ensure the value our contribution creates flows back to the communities that produced it.
We recognize that our holdings (government records, public broadcasting archives, library collections, scientific publications, public-domain texts, openly licensed creative works) are part of the substrate from which AI is built. We commit to making them accessible through formats that enable lawful AI training, described in a shared vocabulary, and recorded in the AI Commons Registry so that downstream use is traceable, auditable, and reciprocal.
A Simple Commitment
The pledge is built around two separable but related commitments. Institutions can sign one without the other; full signatories sign both.
Availability
Make accessible data available for AI training.
Publish your holdings (government data, public broadcaster archives, library collections, scientific publications, public-domain texts) in formats that enable lawful AI training. Document provenance. Provide bulk-access methods that don't require scraping. Where contractual or copyright constraints apply, make the constraints machine-readable.
Reciprocity
Ensure contribution flows back to source communities.
Describe your contributions in a shared vocabulary, so reciprocity obligations travel downstream with them. Record your contributions in the Registry so they are auditable. Where appropriate, attach use-case dimensions that protect source communities from extraction. Contribution is a decision, not an extraction.
The pledge is structured so that different institutions can commit at the level appropriate to their holdings and authority. Each tier has clear, operational obligations.
Governments
Make public data AI-accessible.National statistical offices, government data portals, public records, public broadcasters. Commit to bulk access, documented provenance, and registry coverage. Lead by example in the data your taxpayers already own.
Memory institutions
Open the archives.Libraries, archives, museums, universities. Public-domain holdings, openly-licensed material, institutional collections. Stewards of human knowledge committing to make that knowledge usable for AI in the public interest.
Publishers
Contribute the consented corpus.Where consent exists or can be obtained (open access journals, public-domain catalogs, freely-licensed content) publishers can contribute substantive holdings without contradicting their commercial models. The pledge defines the boundary.
Rightsholder collectives
Establish reciprocity at the source.Collective management organizations, creator unions, content collectives — represent the rightsholders whose contributions need protection. The pledge gives collective voice in defining what reciprocity actually means.
Sign. Convene.Co-shape.
The pledge launched at AI for Good in July 2026. Signatories are being added, and the text is open for input from institutions joining now.
Become a signatory.
Signatories are joining the pledge. Full signatories commit to both availability and reciprocity. Contact us to begin the conversation.
Discuss signing →Host a pledge convening.
Regional convenings, sector-specific working groups, jurisdictional adaptations — the pledge needs local stewards. If you can bring an audience to the table, we provide the framework, briefing material, and follow-up coordination.
Host a convening →Review the text.
Legal counsel, policy advisors, community advocates — the pledge text remains open for input. We welcome institutions joining now who can identify edge cases, jurisdictional concerns, and reciprocity gaps.
Review the text →Underwrite the campaign.
Funder support enables the convenings, the policy briefings, the coordination work that turns a pledge into an actual movement. Contact hello@ai-commons.org for the campaign brief.
hello@ai-commons.org →
SJ Klein
Madisen Taylor
Richard St-Pierre
Amir Banifatemi