OpenAI this month closed its latest eye-popping funding round—one of the largest ever private investments—at the same time as it reportedly further drifted from its nonprofit roots to full for-profit status.
To the Mozilla Foundation, OpenAI’s for-profit evolution is an example of how tech giants and big-money investors can overtake civic-minded mission statements and profit structures in the fast-paced AI arms race.
It’s part of the reason the nonprofit org is pushing a new vision for something it calls “Public AI,” a sphere of initiatives that would give more AI access to nonprofits, government entities, and companies with socially beneficial goals to complement the current commercial space.
Mozilla outlined this new strategy pillar in a report released last month, and it will frame many of the org’s initiatives and advocacy conversations going forward, according to Nik Marda, technical lead for AI governance at Mozilla.
“We’ve been looking at the AI space over the last few years, and seeing both this increased consolidation and closed-ness in the AI ecosystem, and at the same time, this nascent, emerging trend of what we call public AI initiatives, things like governments starting to help increase access to AI resources, and these nonprofit AI labs starting to build open-source AI components,” Marda told Tech Brew. “And so we see that emerging ecosystem as a basis for creating a meaningful public counterpoint to the private AI ecosystem.”
The PBS of AI: The astronomical cost of training and running large language models means that often only the biggest tech companies are able to compete. Even well-funded AI startups have struggled to keep pace with resource demands.
Mozilla’s report lays out examples of public options at each level of development and implementation of AI, from data and compute to applications and accountability tools. One public example was the Biden administration’s National Artificial Intelligence Research Resource pilot, which partners with companies like Microsoft and Nvidia to provide assets like data and computing power to researchers. The report compares its mission to public broadcasters or transit systems in terms of creating an accessible public alternative for a once emergent technology.
Keep up with the innovative tech transforming business
Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.
Marda said his experience as chief of staff of the White House Office of Science and Technology’s tech division, a position he held until late last year, shaped his view of public AI.
“What we thought a lot about in my time at the White House was, ‘How do you use AI for public missions? How do you think about these big national or international ambitions, these big societal problems, and think about deploying resources in a way to tackle those problems?’” Marda said. “So, that framing ended up being really useful in thinking through [the fact that] we’re not doing this just for AI’s sake.”
Policy wishlist: The report also lays out a handful of policy proposals for lawmakers to promote more accessible AI. These include funding open AI infrastructure and research, antitrust action to ensure a competitive industry, financial incentives for public AI, and protections for workers at each level of the AI space.
Beyond those suggestions, the paper lists steps that academics, civil society orgs, tech companies, developers, and the public can take to promote this push as well.
“There’s a role for everyone to play. We’ve seen time and time again that that’s been the case in previous public approaches to technology, and there’s no reason to act like it’s just one stakeholder that can shape public AI, but the public conversation often goes that way, and we’re trying to rebut that here a little bit,” Marda said.