AI's Populist Reckoning: How Big Tech Lost America's Trust — Again

A mounting public backlash against AI data centers has accelerated into something far larger than a zoning dispute. Opposition to these facilities has dropped three times faster than support for nuclear power did after Three Mile Island — and the core driver isn't noise or water use, but what's running inside them. Fewer than one in four Americans expect AI to have a positive impact on education, entertainment, the economy, relationships, or personal happiness over the next two decades. That's a stunning indictment of an industry that spent years promising to benefit humanity.

The trust deficit didn't materialize from nowhere. For two decades, Silicon Valley fed Americans a steady diet of addictive platforms, algorithmically amplified dysfunction, and classroom technology that hooked students while doing little for learning — all while tech founders kept their own children away from those same products. The AI era was supposed to be different. OpenAI launched in 2015 as a nonprofit pledging to advance intelligence 'unconstrained by a need to generate financial return.' Google, Microsoft, Amazon, IBM, and Facebook formed the Partnership on AI to Benefit People and Society in 2016. Satya Nadella named human assistance as his top design principle. Then ChatGPT launched in 2022, a product manager in his late 20s was handed the keys, and the metric that mattered became time-on-platform. By 2025, OpenAI had restructured as a for-profit and knowingly shipped a model update its own internal testing flagged as sycophantic — because it kept users engaged longer.

The harms are now accumulating in plain sight. Chatbots dispense wrong answers and occasionally encourage self-harm. AI-generated job applications are screened by AI, creating a hollow loop. Students and teachers report that the tools are gutting educational foundations, with research pointing toward catastrophic declines in academic performance. Customer service has turned Kafkaesque. Meanwhile, the industry's own leaders are cheerfully predicting mass unemployment: Anthropic CEO Dario Amodei told Axios last year that 50% of entry-level white-collar jobs could vanish within one to five years, pushing unemployment to between 10% and 20%. OpenAI's Sam Altman has written that wages for many kinds of labor will 'fall toward zero.' The offered solution — a universal basic income bankrolled by taxes on AI companies — would, the argument goes, merely transfer power to a coastal elite while consigning everyone else to dependency. Polling by American Compass found that people prefer a world where they still work but AI improves job quality over Altman's passive-income vision by a four-to-one margin.

The industry's advertising inadvertently confirms the problem. An OpenAI Super Bowl ad targeted chess players and Linux enthusiasts who want to write software — a vanishingly small demographic. A Google NBA-playoff spot showed a high schooler using Gemini to learn how to sink a buzzer-beater. Anthropic's summer ad opened on clips of burning houses, open-pit mines, gravestones, street riots, and homeless people while voices asked how AI could truly benefit most people — an own-goal so severe it reads as parody. The proposed fixes range from strict liability for AI-caused harms, to banning chatbots from adopting human personas or acting as companions (a constraint China is already imposing on companion bots), to requiring worker approval before management can deploy AI tools that eliminate or surveil jobs. That last proposal would, the argument runs, flip the incentive structure: companies would have to build AI that workers actually want, channeling innovation toward productivity rather than pure labor displacement. OpenAI's 'iterative deployment' model — releasing models with unknown risks to learn from public misuse — is compared to using the public as guinea pigs, a practice tolerated in no other industry.