Power, Policy, and AI Governance: Taking Agency Over Our AI Future

At WIRED’s Big Interview event, Omidyar Network president, Michele Jawando described how we can all be part of an AI revolution that puts the power back in the hands of people, and why our social identities—not just our data—are at stake.
Image may contain People Person Adult Crowd Chair Furniture Audience Clothing Footwear and Shoe

Faced with JOB displacement, creatives disconnected from the creative process, and waves of nonsensical social media ephemera known as brainrot, it’s easy to feel powerless in the artificial intelligence revolution. But at December 4th’s The Big Interview event, Omidyar Network president Michele Jawando led the audience in a group affirmation: “AI is not destiny, it is design.” That is to say, though it feels like this technology has been foisted upon us, it is the result of millions of small decisions, made one by one, and there is still time to create better outcomes for all of us.

Omidyar Network is a philanthropic organization with a mission to bend the arc of the digital revolution toward shared power, prosperity, and possibility. For Jawando, that means making sure AI is used for more than capitalizing on our labor and ensuring that diverse voices have a seat at the table.

Getting a Seat at the Table

Today’s generative AI models were trained on vast quantities of data—literature, scholarly works, social media discussions—and unleashed upon the world. This raises the question: “Who owns all of that data?” To Jawando, it’s clear that people should own their data, likeness, and creative works, but in these early days of the AI adoption boom, we’ve created and enabled a system that puts the sum of a person’s digital output to work without compensation. “We wanted this technology so fast [that] we made some early decisions,” she said. “And because of that, we are now dealing with the repercussions of those [decisions].”

To change course, Jawando said we have to stop and ask the question: "Who has power, who has recourse, who has the ability to change [that] direction?" Part of the solution is ensuring that a plurality of voices are incorporated into the decision-making process. While her experience leading public policy at a technology giant put her in a position to influence decision-making on a global scale, Jawando emphasized that: "I never forgot that [while] I was able to be in that room and there were so many people who could not. And so I think right now the question is: who's missing from that room before you launch the product? What are the voices that aren't really showing up and what are the perspectives?"

As a result, Omidyar Network puts its effort into developing and supporting civil societies, academics, small NGOs, and startups that provide perspectives that may be overlooked by the hyperscalers but are nonetheless crucial to the conversation. “We spent a lot of time at Omidyar Network thinking about agency… If you can create the conditions so that there are more people who feel like [they] can be a startup creator, [they] can get access to compute and datasets,” more people can be empowered, and society can create a more equitable and diverse innovation landscape.

Fair Use Versus Empowering Creatives

AI’s artistic output started as fairly rudimentary, delightfully glitchy images. In those heady days of 2022, an interesting prompt was just as exciting as the final output. But the latest LLMs (large language models) can ingest a chunk of text and generate increasingly convincing video assets.

While AI may not be autonomously creating capital-A-Art quite yet, the tools are getting better, and in the right hands, they can be used for more than slop. So what does this all mean for creative people—the writers, artists, set builders, and others—who create the culture that AI riffs on to create “new” derivative works? “Art, culture, design, who we are—these are cultural memory,” Jawando said. If we don’t set up structures to pay artists for their work, “we're going to meet a point where [we have to decide if] we only want the slop that comes out, or do we want something more?” To meet this moment, Omidyar Network is currently working with legal experts at Howard Law to start building a modern, non-extractive licensing structure that values human creativity. “I want WIRED to still be here in 40 and 50 years,” Jawando said. “And so we have to think about how to do this… together?”

Governance as a Springboard to Ambition

This is part of a larger conversation around governance. “There are reasons why we make decisions about governance that can change literally the course of how we will meet this moment,” Jawando said. “But you have to be intentional about those choices.” Legislation like California’s Senate Bill 53, which was signed into law this fall, ensures, for instance, that developers of frontier AI models are keeping safety in mind as they run headfirst into the future.

"I think when people hear the word 'governance,' they [think] that means slow down or that means anti-innovation. And again, that is a false binary," she continued. It is once more a question of agency. Jawando contends that to avoid monopolies, echo chambers, and asymmetric technological advancement, we have to consider who has access to capital and opportunity.

For years, the promise of AI was to cure disease or end climate change. While that work is still ongoing, the cultural spotlight is focused on profit. "The ambition is so small," Jawando lamented. "We used to go to the moon! We said, 'We're going to the moon,' and we looked up and we [said], 'We're going to figure it out.'" By decentralizing innovation, we can find new use cases for the transformative power of AI that are decoupled from the fiduciary demands of public companies. To get there, Jawando said we need to open up the playing field through access to public compute—such as the CalCompute provision within SB 53—or access to open-source datasets.

Be the Revolution

When asked what the general public can do to get involved in the AI revolution and retake control of their data, Jawando was optimistic. She noted that because of the proliferation of AI tools, we're all already AI experts and that our voices and opinions are valid. Second, market incentives work: hundreds of thousands of enterprise organizations are choosing to use LLMs that make governance and security a primary concern. "If everybody in here identifies one or two organizations that are doing AI safety research, that are thinking about healthcare and education, and you give them $10 a month, you are making a major difference," she said.

Elevating voices of change—getting more people a seat at a bigger table—can change outcomes for everyone. We can bend the arc of AI toward giving people back ownership of their data, their creativity, and their output. "If we change those narratives," Jawando said, "I think we'll see greater and greater ideas about what's possible."

Learn more about Omidyar Network at omidyar.com