"a political movement and an open-source conspiracy"
The Public AI Network is a coalition working to bring about public AI. Our aims are to ensure public capacity-building is part of the conversation about AI design, policy, and funding; to coordinate related research efforts across the ML community; and to support policy makers and techincal teams implementing public AI.
Many approaches to these goals are being developed today around the world, by a range of public, non-profit, and private organizations. Most are isolated from one another; constrained by access to data, computing, and feedback; dependent on third-party models that they cannot inspect or audit effectively.
Limited public capacity in AI is stifling our collective ability to shape the direction of society. AI is becoming essential infrastructure for information and evaluation, an integral part of the public sector which should reflect the public interest, and the impact of decisions made in training and deploying popular models is growing. Private companies have a head start in capacity, reach, and contribution to public discourse about the use and purpose of AI; coordinated investment of time and energy is needed to bridge that gap.
We are organizing a calendar of events to connect like-minded practitioners; a series of seminars for making progress on specific topics and inviting contributions to publish as collected works; a set of policy papers laying out a framework for advancing and normalizing public AI; and a catalog of shared resources, from compute and participating research groups to a community for finding collaborators, policy guidance and topical expertise.