Checkpoint: How Unions are Taking on AI
Google Deepmind
It’s guzzling up our water to cool the data centers that house its servers. It’s giving our school-age children an anti-intellectual shortcut for homework, tests, and essays. It’s driving up our electric bills with, again, the data centers. It’s giving our judges and lawyers precedent to cite that doesn’t exist. It’s teaching us to be emotionally reliant on a system designed to validate every thought we plug into it. It encouraged a young man in Texas to die by suicide, and it coached him on how to do it.
Turns out, AI is also very bad for workers. In the short term, this new technology sucks up resources and leads to layoffs. In the long term, it undermines human work products and outsources jobs to overseas labor markets. Tech giants — Amazon, Klarna, Oracle, Meta — are laying off workers and giving those jobs to AI technology they’ve invested hundreds of billions of dollars in over the last ten years. So far, though, the tech just isn’t advanced enough to replace an entire human being in an entry-level role, as exemplified by Klarna, which eliminated 700 customer service positions between 2022 and 2024 and replaced the roles with AI chatbots. By 2025, Klarna customer satisfaction had tanked and the company admitted to re-filling the eliminated roles with “remote workers.” Post layoffs, these same companies are realizing the AI isn’t quite up for the job yet, then outsourcing the tasks to workers in other countries to keep margins low. Meta and OpenAI, for example, have quietly rehired workers in Kenya to fill roles previously eliminated for AI. According to Forrester Research’s 2026 Labor Market Predictions, half of workers whose layoffs were attributed to AI will be rehired but replaced by employees overseas at significantly lower pay grades. Because tech companies are laying off workers based on AI’s potential, not its performance, they’re having to employ low-paid workers in the Global South to correct AI’s mistakes.
Imagine you’re a software developer at Amazon. Congrats, you survived your company’s last layoff. Now, the AI that’s being used to replace 10% of your coworkers will also be used to monitor your every move. At Amazon, executives believe that monitoring employees makes them work harder and faster. They’ve employed AI to measure human employee progress, to make sure workers are hitting their productivity targets. Aware is an AI firm whose purported mission is: “To enhance the lives of people and organizations through ethical and sustainable use of emerging technologies.” Aware specializes in monitoring and analyzing employee text, Slack, Zoom, and email interactions, measuring employee sentiments toward their employers and flagging any perceived dissatisfaction with the company. Walmart, Delta, T-Mobile, Chevron, and Starbucks are all using Aware or similar tools to keep tabs on their employees. AI surveillance of all communications makes it much riskier to unionize, proving especially toxic in workplaces like Starbucks that are notorious for union-busting.
“A lot of this becomes thought crime…This is treating people like inventory in a way I’ve not seen.”- Jutta Williams, cofounder of Humane Intelligence
Workers in some parts of the labor market affected by AI are fighting back.
Earlier this year, 2,400 therapists and other mental health professionals in the National Union of Healthcare Workers employed at Kaiser Permanente facilities in Northern California entered a 24-hour strike to protest the use of AI in mental health care. Strikers told NPR reporters that Kaiser has begun using AI systems to triage patients, a duty formerly held by licensed clinicians. Now, patients are triaged via an AI generated in-app questionnaire. The striking healthcare workers expressed concerns over future stability of their jobs, and more urgently, the health and safety of their patients. While Kaiser Permanente reported that they do not use AI to diagnose or make any other medical care decisions, lingering fears among healthcare workers persist. If Kaiser is willing to replace licensed clinicians in the triage phase of healthcare, perhaps therapists will be next.
In July 2025, concession workers at Fenway Park, members of Boston’s Local 26 Hospitality Union, staged a strike against their employer Aramark to demand a new contract with better wages and scheduling, and protections for workers from AI-automated concession kiosks. A new contract was ratified by December 2025 with increased wages and new guidelines for automation. Carlos Aramayo, the president of Local 26, wrote "Aramark worked with our members on staffing and pay, and on how new technology fits the job — so service stays safe, the operation stays strong, and Fenway keeps the experienced workers who make it run." Thanks to the strength of collective bargaining, concession workers who’ve been at Fenway for decades don’t have to worry about losing their jobs to AI.
The Local 226 Culinary Union in Las Vegas gave employers a deadline of November 10, 2023, after which 35,000 of the union’s hospitality workers would go on strike in the largest hospitality worker strike in U.S. history. This strike was narrowly avoided but would have affected 18 casinos and resorts on the Las Vegas Strip. The Culinary Union secured a 5-year contract that brought wage increases and expanded safety protections and technology rights. The provisions around new technology include 6 months of notice before implementing new technology, mandatory negotiations on the implementation of AI, free job training for positions altered or replaced due to automation, and a bonus package that includes six months of health and pension benefits if a union worker is laid off due to AI.
55% of employers who made AI-attributed layoffs already regret those layoffs. Dreams of AI preeminence, where 80% of the workforce has some or all of their labor load replaced by AI, as predicted by OpenAI, are just not materializing. At present, AI’s shortcomings are too vast to replace human workers without yielding a dramatically worsened product. Workers are being discarded and replaced with a fantasy, and the examples of Klarna, Meta, and OpenAI show us that companies will choose AI over their own employees, and will then exploit the labor of people in the Global South to solve the problems that inevitably arise.
The only way for workers to combat the insidious creep of AI into the labor market is with collective bargaining. The at-will workers laid off in droves from tech companies have limited recourse, while those represented by labor unions have the power to advocate for themselves and the longevity of their occupations. Unionization is the best way to fortify against AI expansion. AI tools are fundamentally labor replacing; their goal is to render humans obsolete. Collective bargaining is how we ensure that technology is adopted on our terms rather than brought in to replace us.
