The knowledge thatmade AI possible belongs to everyone.
The infrastructure to govern it does not exist yet. We are building it with the
researchers, institutions, legal advocates, and communities already working
on the pieces.
Who we are
AI Commons is a nonprofit that builds the shared coordination and trust layer
for consent-aware, provenance-rich AI. Every AI system that can write,
reason, translate, or create learned to do so from the recorded work of
human beings – centuries of effort, most of it given freely. We exist to ensure
that the systems built from that shared inheritance remain accountable to the
people they came from, not only to the organizations that built them. We do
not train frontier models. We build the infrastructure that makes AI possible
to govern and trust.
The urgency
Why now
70%+
of widely used training datasets omit licensing information
Aug 2026
EU AI Act enforcement begins.. institutions are not ready
~50%
of major training datasets contain errors in the data they do report
Every institution
deploying AI is duplicating compliance work privately, or not deploying at all
The work
What we're building
Five interconnected pieces of infrastructure that do not exist yet.. and need
to.
1
Provenance and Consent Registry
Machine-readable records of origin, licensing, consent
status, reservations, and transformations so institutions can know what trained the AI they are using,
and prove it to a regulator.
2
AMPL Licensing and Policy Toolkit
Modular contract tools that make permissions between
creators and AI builders explicit, portable, and operational across jurisdictions. No more invisible
pipelines.
3
Compliance Evidence and Evaluation Infrastructure
Reusable manifests, model cards,
benchmark suites, and audit protocols that are built once, shared across institutions, instead of
duplicated at enormous cost by each one separately
4
Reusable Cleared Artifacts
Datasets, multilingual corpora, benchmarks, tools, and reference
models with lineage and terms clear enough to travel across borders, sectors, and use cases.
5
Accountability and Redress
Dispute intake, correction pathways, audit support, versioned
records, and legal architecture, so that when something goes wrong, there is somewhere to go.
Get involved
Who This Is For
FOR FUNDERS
Put capital where the leverage is.
AI infrastructure will be funded one way or
another. The question is where that capital lands.
Registries, legal tools, multilingual corpora, audit
methods. These are roads, not horses. Built
once, used indefinitely.
FOR BUILDERS
Build on a foundation that travels.
Publish lineage. Respect permissions. Contribute
improvements back. The commons gives you
cleared datasets, reusable compliance evidence,
and a shared trust layer, so you are not starting from
zero every time.
FOR GOVERNMENTS & INSTITUTIONS
Require what you can actually verify.
Procurement standards, provenance
requirements, and evidence packs that hold up to
audit and not just brand promises. AI Commons
builds the operational layer that makes
responsible deployment possible at scale.
FOR CREATORS & THE PUBLIC
You are the source.
The knowledge that built these systems came from
all of us- from every person who ever wrote
something down, taught something, made
something, shared something. Which means none
of us are outsiders asking for access.
What We Will Not Do
We think it matters to be clear about what we are not:
— Try to outspend hyperscalers in frontier training.
— Ask institutions or creators to trust invisible pipelines.
— Impose one governance model on every sector or jurisdiction.
Built With and Alongside
Our history
A Decade of Building
AI Commons was founded in 2016 — before most institutions were asking
these questions.
2016
AI Commons is Founded
AI Commons began as a conversation about whether AI would amplify inequalities
or become shared infrastructure. The question kicked off a series of convenings
among institutions who shared a concern: AI would either amplify existing inequalities
or become a transformative resource for collective benefit, depending on how it was
governed.
Dec 2025
AI Commons 2.0 Kickoff: Berkman Klein Center
Harvard University
Over two days, 30+ researchers, builders, archivists, and legal experts
convened to prototype the missing layer of public AI: a consent-aware,
provenance-rich substrate that human institutions have always built for
transformative technologies through libraries, archives, peer review, and
shared public memory. Three interdependent workstreams emerged: data
and provenance standards, legal and policy infrastructure, and shared
training pilots. The goal was not total victory but irreversibility, building
things that matter even if we fail.
Link: aicommons.cc
Jan 2026
From Intelligence to Interdependence
World Economic Forum, DavosThirty years after John Perry Barlow wrote the Declaration of the Independence ofCyberspace, his daughter Amelia Rose Barlow joined AI Commons and the
Cooperative Futures Institute to ask what the next declaration must be. The session,
held at Davos as a “renegade lightning activation,” treated cooperation not as a moral
aspiration but as critical infrastructure: something that must be intentionally designed,
governed, and maintained.Davos was the moment it became part of a longer
declaration.
Feb 20, 2026
Global South Network for Trustworthy AI Launch
India AI Impact Summit, New Delhi
AI Commons joins as one of 15 founding members, launched before delegations from
100+ countries at the first global AI summit hosted by a Global South nation. Regional
hubs launching in South Africa and Brazil.
-> digitalfutureslab.in/news/launch-of-global-south-network-for-trustworthy-ai
Mar 2026
Funding the Commons: Designing Capital for Public AI
Frontier Tower, San Francisco
At Funding the Commons 2026, AI Commons curated and hosted a full-day Funders
Track on Floor 16 of Frontier Tower. The day moved from systems to mechanisms to
experiments, mapping the full spectrum of capital available for commons infrastructure,
from grants and retroactive public goods funding to cooperative structures and
experimental treasury systems. The critical insight that emerged: lack of coordination is
a way bigger problem than lack of capital.
-> Learn more: https://www.fundingthecommons.io/ftc-frontiertower
2016 – present
AI for Good Summits
ITU / UN, Geneva
Nearly a decade of participation in the earliest sustained convenings dedicated to what
AI should be for, not merely what it can do.
Learn more:
https://aiforgood.itu.int/introducing-ai-commons-a-framework-for-collaboration-to-achieve-global-impact/
2018
AI Problem-Solving Framework
An early framework distinguishing AI as competitive capability versus AI as shared infrastructure, shifting the terms of discourse in several institutional contexts.
Read here:
https://docs.google.com/document/d/1jI4cq6febePa4OiZ_He7L5skUaYpHL8EMlcvHVRuR5w/edit?usp=sharing
2018 – 2019
AI Licensing Mechanisms
Among the first attempts to design a licensing architecture for AI built on shared
knowledge. The direct precursor to what we are building now with AMPL.
2019 – 2021
The Global Data Pledge
A multi-stakeholder initiative establishing shared principles and commitments around data governance in the age of AI.
Learn more:
Global data pledge launch – A foundation for tomorrow
2022 – 2024
Scaling Responsible AI · GPAI
Global Partnership on AI
Sustained engagement on what scaling responsible AI requires institutionally clarifying
that the gap between aspiration and practice is primarily an infrastructure gap.
Learn more:
https://www.oecd.org/en/about/programmes/global-partnership-on-artificial-intellige
nce.html
2023
Scaling Responsible AI · GPAI
Hiroshima, Japan
Contribution focused on the infrastructure conditions required to make shared AI
governance principles operational rather than merely aspirational.
Learn more:
https://www.oecd.org/en/publications/g7-hiroshima-process-on-generative-artificial-intelligen
ce-ai_bf3c0c60-en.html
Feb 2026
AI Commons x Khoj · India AI Impact Summit
New Delhi, with Khoj Studios
Part exhibition, part working commons, part long-table salon. AI became legible not
through explanation but through experience.
Learn more:
https://commonkhoj.org/
Mar 2026
From Automation to Flourishing, A Private Salon
Palo Alto, hosted by Mei Lin Fung, VP UN AI for Good
Around Mei Lin’s kitchen table, a small group gathered to discuss the transition of AI
from a tool of automation to an instrument of mass flourishing. The guest of honor was
Kiren Kumar, Deputy Chief Executive of IMDA Singapore and a primary architect of
Singapore’s digital economy, whose work focuses on building the “public rails” that
enable cross-border innovation while ensuring technological advancement leads to
broad-based societal benefit.
Apr 2026
AI Commons x Global South Network: Working Sessions
Virtual
Follow-on sessions exploring how AI Commons frameworks can support Global South
institutions in contributing high-value datasets with clear terms and governance
participation.
Stay connected
Events & latest
RECENT
February, 2026
India AI Impact Summit, New Delhi
Global South Network for Trustworthy AI is Launched
AI Commons joined as one of 15 founding members of the Global South Network for
Trustworthy AI, launched at the India AI Impact Summit, the first global AI summit hosted
by a Global South nation, with delegations from 100+ countries. The Network addresses
a critical gap: Global South institutions steward high-value data in health, climate,
agriculture, and language, but face legal uncertainty and limited influence in global AI
governance. Regional hubs are being established in South Africa and Brazil.
RECENT
March 2026
Frontier TowerSan Francisco
Funding the Commons 2026: Designing Capital for Public AI
How do we fund systems that are meant to belong to everyone? AI Commons curated
and hosted the full-day Funders Track at Funding the Commons 2026, mapping the
spectrum of capital available for commons infrastructure and moving from theory into live
experimentation. The critical insight: the problem is not a lack of capital. It is a lack of
coordination. The day ended with direct matchmaking between funders and builders
working on shared AI infrastructure.
The knowledge that built these systems came from all of us — from every
person who ever wrote something down, taught something, made something,
shared something. Which means none of us are outsiders asking for access.
We are the source
— AI Commons Manifesto, 2026
Work with us
Get in touch
Whether you are a funder, a builder, a researcher, an institution, or someone
who just believes this work matters — we want to hear from you. Tell us who
you are and what you are working on.