Austerity Intelligence

This piece is part of “Ideologies of Control: A Series on Tech Power and Democratic Crisis,” in collaboration with Data & Society. Read more about the series here.

These days, Big Tech oligarchs sail their mega-yachts to private islands with a strong wind gusting at their backs. The Trump Administration awards them lucrative government contracts, and the President appoints one of them to dismantle watchdog regulators and seeks to privatize government services. In addition, the Republican Congress is gutting Medicaid and SNAP (nutrition assistance), which at least 68 percent of Americans will qualify for in their lives, to provide the rich with tax cuts, all while requiring the federal government to buy Big Tech’s AI systems and preventing any state and city from overseeing the technology.

Although American democracy faces constant threats—floods of untraceable money, voter suppression, noncompetitive elections, and rampant disinformation, among others—this moment of collusion between Big Tech, the Trump Administration, and many Republicans in Congress augurs something new. They appear to be employing AI to enforce austerity politics and create permanent instability through decisions that deprive the public of the resources needed to participate meaningfully in democracy. Combined with other harms, like allegedly stealing and using confidential federal data, empowering the administration to target political enemies, and creating false realities, AI is clearing Trump’s path to authoritarian rule.

All forms of democratic participation require something in common: resources. It takes time, energy, and money to vote, contact elected officials, attend meetings, band together with others, imagine better worlds, donate to candidates or causes, talk with journalists, convince others, demonstrate, use the court system, and so on. Unsurprisingly, then, the affluent are far more likely to participate than those with limited means. In a country where nearly 30% of people live in or near poverty and 60% are unable to afford a minimal quality of life, democracy starts at a disadvantage.

Now, AI is exacerbating the divide. Government officials, employers, landlords, and others in power are using AI to make harmful decisions about every key aspect of life.

Insurance companies use AI to deny payment for patients’ necessary medical treatments, and states use it to terminate people from Medicaid or cut home-based care for disabled people., Governments increasingly use AI to determine eligibility for public benefit programs or accuse recipients of fraud. Landlords use AI to screen prospective tenants with often inaccurate background checks, raise rents, and to surveil renters to more easily evict them. Employers use AI to hire and fire workers, to set their schedules and pay, and to monitor everything they do. Principals and law enforcement officers use AI to predict which students might commit a crime in the future and then, in the words of one deputy sheriff, “make their lives miserable until they move or sue.“

In my 12 years as a legal aid lawyer fighting alongside low-income people, I saw the fallout firsthand. Without warning or explanation, my client communities wrongfully lost Medicaid, Unemployment Insurance, or Social Security benefits due to AI. People got terminated when AI sent out incoherent paperwork or failed to recognize valid paperwork that was returned. Benefits with strict asset limits were suspended when AI falsely attributed ownership of bank accounts or property to recipients who didn’t own them. Disabled people with conditions that don’t improve, like quadriplegia or cerebral palsy, got their in-home caregiving services cut because the AI suddenly decided they needed less care than the state’s nurse had determined a year before.

Similar stories of lost benefits, housing, work, and educational opportunities number in the millions. After all, part of AI’s allure is acting at scales and speeds impossible with analog methods. And, because most private companies aren’t required to disclose their use of AI and most governments don’t adequately do so where required, there is no way for people to make sense of what is happening. Plus, many of these decisions can’t be readily challenged. Fickle gods strike—and they can do so again at any time.

The consequences of such arbitrary actions are immediate. People delay medical treatments or incur unaffordable bills, skip meals, spend hours on the phone trying to figure out what happened and what options are available, wait in line at food banks, beg with landlords or power companies not to evict them or turn off the lights, and plead with employers for flexibility because their usual transportation or childcare options are now unavailable. The intensive time, physical toll, and stress involved don’t quickly end and recur with every new harm.

Made to scrape even harder for survival, how are low-income people able to attend meetings, stand in ever-longer voting lines, or activate around issues in a sustained way? And, if politics doesn’t produce justice, will they believe such efforts matter?

Excluding low-income people from democratic participation diminishes the possibilities for a better world. Low-income people constitute a sizable bloc of voters—enough to shift elections, even if they vote at lower rates than their more affluent peers. More broadly, though, low-income people, often in alliance with others, have driven movements for fundamental economic rights that the broader public enjoys, including workers’ rights to unionize, minimum and overtime wages, and workplace safety, civil rights and anti-discrimination laws, public benefits programs, and community-based living for people with disabilities. Entrenched inequality makes us all poorer.

While low-income people are the first targets of actors using AI to make decisions, they won’t be the last. Employers already use AI to decide the fates of higher-income workers such as finance executives, therapists, insurance underwriters, radiologists, and computer programmers. Many of the insurance companies’ and landlords’ practices mentioned extend to people at all income levels. Meanwhile, the Trump administration has already made universal programs like Social Security harder to access and is threatening to automate away the federal government workforce, a bedrock of middle-class jobs. Less immediately, long-term reliance on AI hurts our collective well-being by further degrading the environment, consuming scarce water, and raising energy prices.

While nobody will escape AI, only a few will prosper under it, as its track record so far is decidedly one-sided. There are no examples of AI being used to meaningfully improve access to jobs, housing, health care, education, or public benefits at a scale that matches the scope of its harms. This current dynamic suggests that the technology’s underlying purpose is to entrench inequality and reinforce existing power dynamics.

How, then, do we invigorate democracy when AI is being used in ways to undermine it?

First, we help people survive it. Local community organizers, legal aid attorneys, and frontline social service providers can help to figure out what’s happening, fight back, and connect people to whatever support or mutual aid exists.

Second, we build power among the people most directly harmed. AI’s rampant injustices can coalesce people devoted to different issues, including the safety net, workers’ rights, housing, the environment, criminalization of Black and Brown communities, education, corporate power, and democracy reform.

Third, we relentlessly posit AI as an illegitimate way to make key life decisions. It is opaque, irrational, inaccurate, deceptive, and often based on junk science. In contrast, when human decision-makers cause harm, make mistakes, or act with bias, they are generally easier to challenge and hold accountable.

Fourth, we focus on a long-term, positive vision of a robust, participatory democracy rooted in an egalitarian society that guarantees core human rights and incorporates technology to further such ends.

Victory won’t come easily. But the Trump Administration’s open embrace of Big Tech and AI creates an urgent and visible crisis, one that could spark the kind of broad-based resistance needed for change. Indeed, this technological reckoning may be the only path to democratic revival.