What We Covered

This webinar brought together California voices around a shared premise: whether AI deepens existing inequities or is bent toward economic justice depends on our political and economic systems, not on the technology itself.

Rebeca Rangel of the Asset Funders Network opened by situating California as the epicenter of AI—home to the frontier companies, hundreds of active data centers, more billionaires than any other state with collective wealth over $2 trillion, and one of the highest poverty rates in the country, with 7 million residents lacking the resources to meet basic needs. She offered a map of three layers where AI outcomes get decided: where we feel it, on the job and at home and in our neighborhoods; what supports us, the systems we rely on when work and income change; and who sets the rules, meaning ownership, taxation, and market power. Whatever a funder’s issue area, she said, there is a lever on that map that already belongs to their portfolio. She borrowed Omidyar Network’s framing: regardless of what you major in, you need to minor in AI at a minimum.

Elaine Chang of the Institute on Race, Power and Political Economy at The New School opened with why there is still time to influence how this technology gets absorbed. The work that will actually shape AI’s impact, she said, mostly isn’t labeled AI policy—it is labor law, tax policy, and resource distribution. She compared call centers in the United States and Germany, where similar voice recognition technology produces opposite outcomes: narrow tasks, high turnover, and unpredictable hours domestically, and one worker handling the whole interaction with more discretion in Germany. She cited MIT research finding that two-thirds of wage inequality since 1999 stems not from occupation but from whether a worker lands in a firm that captures surplus. Her prescription is to unknot those incentives through tax treatment of labor relative to capital, worker classification, and protections against surveillance wages, and she pointed to California’s AB 1340 as recent momentum. More broadly, the Institute’s human rights economy frame calls for guaranteed resources and real agency, because the incentive to make a job cheap and disposable only works if the person in it has nowhere to go.

Teri Olle of Economic Security California described AI as a new chapter in an old book on an old shelf, cautioning that being normalized is not the same as being harmless. She asked the audience to hold two frames throughout. The first was who decides how this technology gets used, and will it deliver the broad-based prosperity so often promised or keep concentrating money, wealth, and power in the hands of a few? The second was that policymaking is at best responsive and almost never precautionary, which puts everyone working toward a better outcome at a structural disadvantage, running to catch up. Against that backdrop, she offered four harms, none of them invented by AI but all of them supercharged by it in scope, speed, and scale. Algorithmic price fixing, where subscription services scrape data and push recommended prices to landlords, hotels, and gas station retailers. Surveillance pricing, which sets individualized prices based on location, phone battery level, or the hour of the night. Data centers whose electricity demand tripled between 2020 and 2024, with costs landing on local ratepayers even as the facilities receive tax breaks. And harm to small businesses, illustrated by a Palm Springs shop owner who deliberately chose not to sell on Amazon and discovered her entire catalog scraped and posted there without her knowledge or consent.

Janey Rountree of the California Policy Lab at UCLA described CPL’s early warning system, which tracks which jobs are being displaced, in what industries, in what geographies, and among which workers. The finding so far is no large impacts statewide, though there are pockets, particularly among highly educated workers in the Bay Area. The value of a tracker, she noted, is rarely in what it shows at the start; it is in the change over time. She also walked through the lifecycle of layoffs, from prevention while people are still employed, to cash assistance while they are unemployed, to workforce reentry and wraparound supports, singling out short-time compensation, which lets an employer cut hours while maintaining benefits rather than laying people off and is known to only about 2 percent of California employers. She then offered CPL’s homelessness work as an example of AI built for public benefit: a predictive model that flags people at high risk among LA County health and mental health clients, so the county can reach out and offer cash assistance and case management. She noted in the chat this model worked because it was co-developed over years with social workers and clinicians rather than built on a virtual machine with a pile of data, and that interdisciplinary capacity, she said, is as much the call for innovation as any particular idea.

Lul Tesfai of the James Irvine Foundation described AI as a cross-portfolio issue rather than a standalone program, anchored in Irvine’s North Star that workers paid low wages in California have the power to advance economically. The foundation has been attentive to how technological change, and automation in particular, shapes the outcomes of workers paid low wages. In 2023, prompted by the rapid adoption of generative AI, Irvine began asking what happens when disruption reaches workers paid low wages, who have less financial cushion, less bargaining power, and less ability to withstand an economic shock. The exploration has since widened well beyond job loss, to job quality, algorithmic management and productivity expectations, worker voice, how productivity gains are shared, and how AI is applied outside the workplace in ways that undermine economic security. She closed by noting that this is a moment of extreme uncertainty, but there is a great deal we already know and can act on. Now is the moment to invest in power, protections, and the institutions through which people, particularly those most marginalized, shape what happens next.

What The Panel Recommends Funding

None of what follows requires an AI portfolio or a technology background.

Fund policy advocacy and organizing. Addressing who controls how the technology gets used takes organizations with the capacity to develop policy and build guardrails, funded as sustained structural work rather than discrete projects. Organizing works, and it often works fastest close to home. Siting decisions on data centers get made first at city council and planning commission meetings, and it was that local pressure that moved the state, with seven data center bills passing in the California legislature this session.

Invest in support systems. When workers have genuine agency and real alternatives, the terms of work shift. Some supports act as enablers, letting people take risks and retrain: cash during retraining, apprenticeship and earn-and-learn models, capital to start a business, affordable access to bandwidth and AI tools. Others act as a floor, so no one is trapped in a job they cannot leave: guaranteed income and portable benefits. Expanding these alternatives and building worker power is itself a lever on how AI gets used.

Fund the domain expertise, not just the technology. Two public-benefit uses of AI discussed in the session share a pattern of centering those most affected and the organizations serving them. The homelessness prevention model was co-developed over years with social workers and clinicians. Ask Aya, which gives domestic workers guidance for real situations at work, was built with the collective wisdom of workers themselves and on the National Domestic Workers Alliance’s long history of organizing them. Building that kind of relationship requires sustained investments.

Fund AI upskilling. Philanthropy’s distinct advantage is that it can test things before there is evidence, and then make the case for public systems to scale what works. AI upskilling is one opening, since it is not yet well understood and is ripe for testing community-based programs that reach people currently left out of our economy.

Resources shared during the conversation

Empower or Exploit was co-hosted by Asset Funders Network California and Philanthropy California.