And invest in the good kinds

The mounting evidence overwhelmingly points to Big Tech’s AI efforts causing much more harm than good. The general American populace agrees. A recent Pew Research report tells us that more than half of Americans rate the societal risks of AI as high, Americans are much more concerned than excited about the increased use of AI in daily life, and a majority want more control over how AI is used in their lives.

Thankfully, people are starting to wake up and push back against the notion that Big Tech AI is here to stay and there’s nothing we can do to shape or change it. We still have time to demand parameters for AI’s role in our lives, and to shape what those parameters look like. But there is far less accessible information highlighting how people are successfully changing and shaping it for the better, or simply pushing back. You wouldn’t know from most AI media coverage all the myriad ways people are doing so. Here’s shining a light on a few.

Stopping data centers – success is possible

The biggest and most impactful way that communities are pushing back against this undemocratic AI gold rush is by resisting data centers. A powerful data center resistance movement is growing not just across the United States, but also around the globe.

It’s a real David vs Goliath situation, with new, giant Big Tech data center projects being approved all the time and everywhere, and communities often left to respond in whack-a-mole fashion only after the secret deals are done. But there is a growing decentralized social movement that is unique in some promising ways. For one, it’s becoming abundantly clear to residents all over that data centers are bad for them. To residents, Big Tech’s lies aren’t holding water (but their data centers sure do use a lot of water!). By now, there have been enough examples and experiences for people to know the true effects of these data centers coming to their neighborhoods. They can see through the shiny, shallow stories of promised economic development. Residents are finding it easy to connect the dots between energy price hikes and data center development. Especially in desert communitiesand places with strained infrastructure already, there is visceral residential concern and resentment against these resource-guzzling facilities.

Further, the movement stands to benefit greatly from the power of diversity and rare bipartisanship. In a review of elected officials’ statements, 55% of politicians who publicly opposed data centers were Republicans. There are so many reasons to be opposed to data centers. Whether motivated by environmental impacts, energy grid strain, residential price hikes, tax incentives, noise pollution, jobs, resource consumption, concern over property value, or anger at a lack of transparency, it doesn’t matter. Together, diverse stakeholders are able to sound a unified voice of opposition against data centers.

Big Tech is targeting its data centers in communities they bet against having the political power to fight back. Largely concentrated in the South, those communities are often the same ones who have already been targeted time and again, many times subject to environmental racism. But they are no strangers to fighting back.

Beyond just the South, a decentralized movement across the country is coalescing. The networks of people resisting data centers is diverse and expansive, from residents to grassroots groups, activists, nonprofits, national organizations, environmental groups, policy and research groups, racial justice organizations, homeowner association groups, and legal advocates. Coordinating entity Data Center Reform Coalition was founded in 2023 to help align these disparate efforts. According to their recent report, there are at least 142 activist groups across 24 states organizing to resist data centers.

Their range of strategies include:

Spreading information through social media, gathering support through online petitions, educating through word of mouth, door-knocking, and community teach-ins. Data Center Watch has tracked over 23 petitions with more than 31,000 signatures against data center development since 2022.
Voicing concerns through letter writing campaigns, public comment and visible shows of activism at government meetings.
Demanding accountability through filing lawsuits and open-records requests, and calling for environmental review, resource use transparency, and disclosure of nondisclosure agreements that local officials often sign.
Blocking and delaying projects through environmental impact complaints, leveraging local ordinances, coordinating with local officials to push for moratoriums, and organizing ballot initiatives.
Negotiating for benefits through ensuring local residents stand to tangibly benefit from data centers in certain ways, such as through community benefits agreements (CBAs) that ensure long-term, protected jobs.

Coalition-building organizations like MediaJustice are taking on the role of bridging and sharing resources like this organizer toolkit to build power across these varying entities, united for the same goal.

The fight against centralized, Big Tech AI can feel futile in its scope, but since permitting decisions happen at the local level, there is so much power in acting locally. These data centers are getting approved in very sneaky, undemocratic ways. Local officials are falling for the AI hype these Big Tech companies spew. City officials are turning their back on the public and signing nondisclosure agreements for these companies. Local communities, residents and tribes are getting left out of these major decisions entirely. But the power of organizing and demanding democratic processes work for the communities they were set up for, does still work. It’s more important now than ever to plug in and engage in democracy at your local level. The way to fight oppressive, fascist, top-down systems is from the bottom up. Collective action works.

Activism success stories

According to Data Center Watch, 16 data center projects have been delayed (10) or stopped (6) across the US, from spring 2024 to spring 2025, thanks to community pushback. These projects amount to roughly $64 billion in US data‑center projects so far avoided.

These are additional wins from this year (not included in Data Center Watch’s analysis):

In College Station, Texas, the city council unanimously rejected the land sale for a proposed data center after nearly six hours of testimony from concerned residents.
In Jerome Township, Ohio, trustees enacted a unanimously-passed nine-month moratorium on new data center construction to allow for time to evaluate effects of existing facilities on residents.
In response to data center expansions in South Fulton, Georgia, the Black Voters Matter Fund has been educating residents left out of zoning discussions, as told to Capital B News.

In Chesapeake, Virginia this summer, Chesapeake City Council unanimously rejected what would have been the region’s first major data center, after concerned residents flooded city hall with two and a half hours of public comments.
In Tucson, Arizona, community pressure resulted in the Tucson City Council’s unanimous vote against building the massive Amazon data center, Project Blue.
In Indianapolis, Indiana, Google withdrew its data center proposal in light of considerable public opposition.
In St. Charles, Missouri, public opposition to a proposed data center pushed the developer to withdraw its application for a permit.
In Prince George’s County, Maryland, the County Executive has issued an executive order to pause data center development for six months in response to a hundreds of residents-strong protest and a petition signed by 20,000 against data center plans.

Not every instance of community pushback leads to the end of a data center. But canceling data center projects is extremely unlikely to happen without robust community pushback.

Making them pay

Big Tech companies also choose locations for their data centers largely based on where they can get the biggest tax breaks and the cheapest electricity rates for industry. In 2022 alone, data centers received tax breaks worth $180 million. Meanwhile, electricity rates for residents are soaring as a result of these data centers. That’s why another resistance strategy utilized by some organizers and lawmakers is trying to make these companies pay more of their share.

This summer, the Oregon state legislature passed a bill addressing just that. In a state where data centers consume 11% of its power, the bill would create a new classification for data centers and other large industrial energy users to pay their share of electricity use and costs.

Just last month in California, Governor Newsom signed into law SB57, which will require the Public Utilities Commission to assess the extent of extra electrical costs on consumers as a result of data centers.

American Electric Power (AEP) Ohio has been pushing to require Big Tech companies to pay more than other commercial infrastructure due to the excessive resource strain of these data centers, and these companies’ obvious ability to afford it.

Pushing for regulation

According to a recent University of Maryland survey, large majorities of Americans are concerned about unregulated AI and want AI regulation that guides its development.

So far, the US has been allowing Big Tech companies to self-regulate their AI development (we see how that’s going). Still no federal legislation has been able to pass, BUT there are now 131 enacted general AI state laws, and 52 adopted state laws. This year, all 50 US states, plus Puerto Rico, the Virgin Islands, and Washington, D.C. have introduced AI legislation. That may seem like a lot, but the enacted laws from this year comprise just 10% of the more than 1000 bills that were introduced. This report by KPMG organizes US state laws and regulations into two categories, comprehensive and targeted AI legislation. ‘Comprehensive’ AI laws intend to impose requirements on both developers and deployers of ‘high-risk’ AI systems, but only one of twenty introduced comprehensive AI bills has been enacted. Given how far-reaching and nuanced the dangers of different AI technologies are, many, many laws are needed to address the numerous risks appropriately.

We also desperately need overarching federal regulation, something akin to the EU’s A.I. Act, which classifies AI technologies according to their risk level, and prohibits AI activities categorized as ‘unacceptable risk’ and places transparency requirements on the companies that create massive general-purpose AI tools, and fines companies that violate regulations. The federal AI LEAD Act has been introduced and would allow anyone in the US to sue Big Tech AI companies and hold them liable for any harm caused by their products. Unlike to pass, it would be a huge step in making AI products safer. We should keep pushing for that, but since it is much more out of reach during this federal political landscape, it serves us to double-down on the patchwork, state-based approach for the time being.

Demand an international safety agreement

Another arguably imperative method of regulating AI is through an international safety agreement of some sort. There is precedent for such a thing, including a Nuclear Nonproliferation Treaty and a Biological Weapons Convention. For such a powerful, global technology, there should be a corresponding unified global response and stance. This could look like the creation of an international agencysuch as the International Atomic Energy Agency, which was created to limit the spread of nuclear weapons. So far much of the Big Tech-disseminated discourse around AI has been in terms of global competition, but it doesn’t have to be that way. We can shift the cultural narrative to one of collaboration and reigning in these frightening technologies together, for the protection of all.

Make your voice heard (irl)

The positive side of AI now touching nearly every arena of our lives is that we now have power to shape it in nearly every arena of our lives. In all the different sectors of life we participate in, we can make our opinions known on the use of AI technologies there.

That includes every axis of society – schools, doctors offices, and of course, workplaces.

Workers are organizing. Here are some examples:

Amazon employees just released this open letter demanding a more responsible AI rollout. If you’re not an Amazon employee, you can still sign their Solidarity Statement.
Pennsylvania’s social service employees union, SEIU, established a workers council on AI.
The American Association of University Professors (AAUP) at Miami successfully pressured management to work collaboratively with them on addressing AI.
The National Nurses United union introduced a bill of rights regarding the use of AI in healthcare.
And of course, the successful Screen Actors Guild and Writers Guild of America sustained strike and negotiations of 2023.

If you can unionize your workplace, do so! You can engage in bargaining anti-AI clauses into union contracts. An especially incredible impact could be achieved if workers at all these Big Tech companies unionized. Tech workers have been gradually shifting toward unionization in the past few years, but it is an ongoing battle involving a lot of pushback from these companies. Workers at these Big Tech companies can and should also organize to pressure their companies to bring on ethicists and philosophers to guide every big decision in the AI development realm. Applied philosopher Cansu Canca argues back in 2020 that tech engineers, though well-meaning, do not have the expertise to formulate specific ethical standards to hold their employers to. Thus, without expert ethical guidance, any demands for ethical AI result only in vague ‘ethics washing’ that cannot be operationalized. These experts should be collaborating throughout all the phases of AI development and implementation. If Big Tech’s AI tools were developed thoughtfully, carefully, and collaboratively this way, it would surely eliminate a lot of need for subsequent regulation, too.

How to support good AI

In addition to fighting to resist and shape Big Tech’s AI moves, it’s crucial to spend energy lifting up and supporting the good kinds of AI, too. The responsible narrow AI tools built to solve real, discrete problems in the world, and the organizations building better approaches for the field at large. Supporting the good kinds of AI development sends a message to the broader AI industry as a whole of what we actually want AI to do and be capable of. The more diverse makers of helpful, targeted AI tools there are, the more balanced and democratic the industry can become. And the less power will be concentrated in the hands of a few billionaire bad actors.

Here are some organizations doing good AI work: