The army that is quietly reshaping your future arrived in Washington wearing suits, not uniforms. It does not carry guns. It carries checkbooks. And in the past three years, its ranks have grown larger and faster than almost any invasion force the capital has ever seen.
More than 3,500 federal lobbyists — one out of every four working the corridors of power — reported working on artificial intelligence issues in 2025, according to a new report from Public Citizen. That represents a 170 percent increase in AI lobbyists since 2022. The number of lobbyists specifically targeting data centers, the physical engine rooms of the AI revolution, exploded by 500 percent over the same period.
Here is what that means in plain English: For every three lobbyists trying to influence your government on health care, taxes or the federal budget, roughly one lobbyist is pushing an AI agenda.
More people are paid to shape AI policy in Washington than are paid to shape defense policy. More than energy policy. More than nearly anything except the budget itself. Only federal budget, tax, and health care issues attracted more lobbying activity in 2025 than AI. An entire industry has materialized overnight, and it has brought plenty of cash.
The U.S. Chamber of Commerce, that venerable voice of American business, deployed the largest AI lobbying force in 2025, with 91 people registered to press the flesh and the case for machine intelligence. Microsoft sent 63. Meta sent 55. Intuit, the tax software giant, sent 51. Amazon sent 48.
Every single one of the top 30 lobbying entities identified in the report, and 91 of the top 100, represent either corporations or corporate trade associations. Not a single public interest group cracks the top 30. Not one.
The push is broad because the technology is broad. Two hundred companies in the software and services industry accounted for 30 percent of all AI lobbyists. But the rest came from health care, media, finance, manufacturing, and defense — a testament to how thoroughly this technology has penetrated every corner of the economy.
Here is what the AI industry wants: Congress to pass a federal law that overrides any state effort to regulate the technology. The industry wants what is called “blanket preemption” — one rule for the whole country, drafted by those who stand to profit most from its laxity. It wants “regulatory sandboxes” where companies can test AI systems without oversight. It does not want a new federal AI regulator. It does not want enforceable guardrails.
This is not speculation. The White House released a National Policy Framework for Artificial Intelligence in March 2026 that explicitly calls for preempting state laws, creating regulatory sandboxes and establishing no new federal regulator. The framework’s legislative recommendations read like a corporate wish list because they were written with substantial corporate input. That is what you get when you field an army of 3,500 lobbyists.
Consider the mechanics of this influence campaign. AI-active firms in the health care and finance sectors alone reported a combined $503 million in total lobbying expenditures in 2025. OpenAI spent roughly $2.2 million on federal lobbying in the first quarter of 2025 — more than the entire independent AI safety research field received in grant funding during the same period.
Mike Tanglis, a research director for Public Citizen and co-author of the report, does not mince words. “AI is already causing significant harm to children, our energy grid, our environment and the job market,” he said. “Leaving a society-transforming technology completely unregulated at the federal level is already having grave consequences.”
He is not talking about some distant science fiction future. He is talking about right now.
Consider the evidence already in plain sight. In 2025, employers cited AI as the direct reason for at least 55,000 job cuts. The consulting firm Challenger, Gray & Christmas has described AI’s impact on the labor market as akin to a “tsunami.” Dario Amodei, the CEO of Anthropic, has warned that AI could eliminate half of all entry-level white-collar jobs within the next five years, potentially spiking unemployment to 10 to 20 percent. That is not a distant hypothetical. That is the projection from one of the industry’s own leaders.
The environmental toll is staggering. AI systems generated an estimated 80 million tons of carbon dioxide in 2025 — equivalent to the annual emissions of New York City. They consumed up to 765 billion liters of water, surpassing the entire global bottled water industry’s annual demand.
Google’s greenhouse gas emissions have risen 48 percent since 2019, largely due to AI-driven data center expansion. Microsoft’s emissions have grown 29 percent since 2020. Both companies have abandoned or significantly modified their carbon neutrality commitments.
Deepfakes are already influencing elections. In 2025, a study published in the Journal of Creative Communications found that people struggle to identify deepfake videos and that their opinions are affected by this type of misinformation. Republicans appear to be using the technology more frequently than Democrats in the current election cycle, but that disparity is unlikely to last. The tools are getting cheaper and more convincing by the month.
You might wonder why you have not heard more about this. Part of the answer lies in who owns the outlets that would normally sound the alarm. Five companies now control 90 percent of American mass media, compared to 50 companies just 30 years ago. The Roosevelt Institute has documented how this concentration makes the entire information ecosystem more vulnerable to state pressure and corporate capture. Five companies. Ninety percent control.
Sen. Bernie Sanders warned in November that the concentration of media power in the hands of a few ultra-wealthy individuals is “common in authoritarian societies.” He called for action against what he described as oligarchic control of American information.
The Free Press Media Capitulation Index ranks the 35 largest commercial media and tech companies on a scale of independence — and found that only two, Bloomberg and Netflix, earned a full star. The rest, in the index’s telling, have compromised their independence in exchange for political favors or to get the Trump administration off their backs.
This is the context in which 3,500 lobbyists operate. An industry with virtually unlimited resources is pushing for virtually no rules. And the institutions that would normally expose that effort are themselves increasingly owned by people with a vested interest in the status quo.
Eileen O’Grady, a researcher for Public Citizen and co-author of the report, sees a pattern that Americans have lived through before. “AI is at severe risk of becoming the next chapter in the story of unchecked corporate power that causes widespread and avoidable economic devastation,” she said.
Her colleague Mike Tanglis put it even more bluntly. “Wall Street and Silicon Valley lobbyists insisted that the public just couldn’t understand their industry and therefore, we must leave it to them to regulate themselves. That didn’t work out well.”
He meant the 2008 financial collapse. He meant the opioid crisis. He meant the decades of deception by the tobacco industry.
In each case, industry lobbyists promised their technology was too complex for outsiders to understand, they promised that self-regulation would suffice, and they promised that any harm would be swiftly corrected.
In each case, they were wrong, and the public paid the price.
O’Grady offered an alternative. She said Congress should reject blanket preemption and deregulatory sandboxes. It should refuse to bail out speculative AI finance. It should build enforceable, sector-specific guardrails. It should invest in regulators’ technical capacity and enforcement capabilities. And it should prioritize protections for workers, children, and communities.
None of this is radical. It is simply the application of common sense to a technology that is already changing everything. The question is whether Congress will listen to the 3,500 lobbyists — or to the 350 million Americans who will live with the consequences.
Tanglis warned that if the industry has its way, the lobbyists will decide how to implement safety protocols, when to alert authorities about problems, and when to prioritize public interest over profits. “We’ve done this before,” he said.
We have indeed. And the scars are still visible.