Rein in Big Tech
Designing a Public AI Workshop
10. 02. 2024
Experts from diverse fields convened in D.C. to advance Public AI infrastructure, emphasizing public interest, policy, and innovation in AI's future
Experts from across the fields of policy design, computer science, economics, civil society, academia, industry, philanthropy, and public service converged in Washington, D.C. to learn from one another, build new relationships, and advance the idea of building public infrastructure for AI in the United States. Approximately 30 participants shared their work, assessed existing progress toward building Public AI, and explored how we might bend the arc of this technology toward a vision of broad-based prosperity. Building on their governing AI work, Vanderbilt Policy Accelerator plans to publish a report coming out of the workshop that analyzes global and domestic approaches to Public AI to help identify successful strategies and inform future initiatives.
“This isn’t just about this particular moment in AI. It’s about saying we’re going to engage in the digital economy in a very different way than we have done in previous years. That’s a huge shift and an important first step,” said Arati Prabhakar, Director of the White House Office of Science and Technology Policy and Assistant to the President for Science and Technology. “Public AI is so important because you are talking about exactly what needs to happen in government and in other sectors as well. Because this will be something that all of society is gonna have to surge to do.”
Presenters from Public AI projects underway around the world laid out some of the trials and triumphs of their designs thus far, touching on everything from procurement plans for GPUs, to national security concerns over inevitable investment in foreign companies. Using the framework from Mozilla’s recent Public AI paper and Vanderbilt Policy Accelerator’s brief on antimonopoly tools for governing AI, we assessed different types of investments, projects, and governance structures that could center public interest while promoting innovation across the stack. Knowing that monopolies and oligopolies already exist along the stack for AI development, we debated points of intervention where public investment could transform the market, and explored outcomes that are in the public interest.
The private sector is pouring tremendous amounts of resources into AI development. There is a real risk of not dreaming big enough at this early stage to position the public sector as a strong counterpoint to leverage this emergent technology for the public good. We have a real opportunity to decide how this technology, and therefore, our economy, takes form.
Together, the group wrestled with important questions like:
- What is the highest value infrastructure to focus on if our aim is to create meaningful public alternatives for AI?
- How do we make sure that we aren’t subsidizing what the private sector will already do?
- What are the technical, talent, and capacity opportunities and limitations for Public AI?
- What can we learn from research and writing related to Public AI, and how do we mobilize these findings into concrete action?
- How do we build resiliency into Public AI infrastructure so that it can withstand changes in the design, needs, and inputs for AI systems in the future?
- Can we find policy solutions that have the desired antimonopoly market effects while simultaneously centering human needs and experiences?
We walked away with a clear understanding that there is a critical need to establish Public AI initiatives that ensure technology serves the public interest rather than solely private profits.