Unions Lead the Fight for Human Oversight in AI

The future of work, once imagined as a gradual evolution, now appears to be accelerating. Whether this transformation leads to widespread unemployment or to new opportunities will depend largely on how governments, companies, and unions respond to the disruptive power of AI.

The convergence of corporate warnings, union demands, and bipartisan political will suggests that the United States may be on the verge of establishing a national framework for AI regulation. While states like California are leading the way, the challenge lies in translating these efforts into federal law that balances innovation with safety and worker protection.

Since the 1960s, automation has steadily reshaped industries across the globe, often resulting in workforce reductions and the elimination of routine jobs. Assembly lines, logistics, and even clerical work have all been transformed by machines and software designed to cut costs and boost efficiency.

Now, a new frontier is emerging: what experts call “physical AI.” This refers to the integration of artificial intelligence into robotics, creating machines capable of performing a wide range of tasks once reserved for humans. Unlike traditional automation, which focused on repetitive mechanical functions, physical AI combines advanced algorithms with robotics, paving the way for intelligent systems that can adapt, learn—and replace workers in diverse sectors.

Having a Human behind the Wheel Is Essential 

Workers across industries justifiably fear being displaced by artificial intelligence, regardless of their specialization. The alarm first sounded in Hollywood in 2023, where screenwriters and authors launched strikes to protest the use of AI in scriptwriting and creative production. That movement revealed that automation was no longer confined to routine or manual labor; it was encroaching on creative professions once thought immune to technological replacement.

Having a human behind the wheel is essential for public safety.

Since then, the concern has spread to literature, education, healthcare, transportation, retail, and other service sectors, fueling a broader wave of resistance. Employees and unions are demanding regulation, transparency, and participation in decisions about AI implementation.

“What we don’t want is for AI to replace the hardworking people who make our workplaces run,” Robert Gonzalez, communications coordinator of Teamsters Local 1932, said to NPQ in an interview. Gonzalez stressed that if artificial intelligence is going to be introduced into the workplace, unions insist there must be clear safeguards in place to protect workers.

“With Teamsters, we also represent truck drivers. In California, there have been efforts to introduce automated vehicles, which would mean replacing human drivers with robots. That is dangerous, and it is fundamentally a public safety issue,” he said. According to Gonzalez, the Teamsters are suing the California Department of Motor Vehicles (DMV) over its implementation of AI in trucks, stressing that having a human behind the wheel is essential for public safety. Gonzalez added that the union is making progress in this fight and intends to continue moving forward.

If AI systems begin to replace these kinds of roles en masse, the social and economic consequences could be severe. Beyond job loss, there is the strong risk of de-skilling, where workers lose expertise because machines have taken over the tasks that once required human judgment.

According to a 2026 article in Frontiers in Medicine, in modern terms, de-skilling is “the gradual erosion of independent clinical reasoning skills, together with crucial elements of clinical competence.” In the present moment, de-skilling “represents a pattern of dependence on technology, especially Artificial Intelligence (AI).”

Being Organized as a Long-Term Strategy

According to a poll conducted by the AFL-CIO, nine out of 10 workers favor measures that ensure human oversight in employment decisions, transparency in how AI is used, and strong guardrails against applications of AI in the workplace.

The survey of 1,588 respondents across the country found that 95 percent want a human to be the final decision-maker in matters impacting workers, while 92 percent support strict safeguards and accountability for employers deploying AI. Union leaders highlighted recent victories in collective bargaining agreements, such as protections won by media workers and nurses, ensuring that AI cannot be implemented without union approval or in ways that undermine professional judgment.

AI can never teach a child the way a human being can, since teaching is rooted in empathy and connection rather than mere information delivery.

Despite this strong support for guardrails and oversight, only seven percent of workers said their employers disclosed how AI monitors their work, underscoring the lack of transparency. When asked who they trust most to protect them from AI, workers placed unions far ahead of political parties or employers, reinforcing the AFL-CIO’s claim that the labor movement is the most reliable defender of workers in the age of artificial intelligence.

Vicki Kurzydlo, a prominent union leader and veteran educator with over 32 years of experience in Chicago Public Schools (CPS), currently serving as the recording secretary for the Chicago Teachers Union (CTU), emphasized in an interview with NPQ the importance of reframing the discussion around artificial intelligence in education. “One way to approach the issue is to change the way we talk about AI. It’s not simply whether AI will displace educators; it’s about what we lose when we remove the human element out of education, and out of any labor, really,” she said.

Kurzydlo explained that unions must organize before decisions are made for them, demanding transparency about what technologies are introduced, what information they collect, how that data is used, and how workers can challenge those decisions. Kurzydlo also warned about the risks of surveillance and constant monitoring in the workplace. She highlighted that the CTU has historically fought to improve the school day for students—winning contract provisions such as recess, smaller class sizes, more nurses, and additional paraprofessionals—and now seeks to establish safeguards as AI enters the classroom.

“Right now, we’re working with groups of teachers within the union. There’s an active AI working group that not only studies and reviews policies but also develops new ones—because, at this point, there’s nothing set in stone nationally,” she said. The union has introduced an AI resolution establishing initial guardrails, marking an important beginning. Teachers are joining forces to resist AI collectively, underscoring that AI can never teach a child the way a human being can, since teaching is rooted in empathy and connection rather than mere information delivery.

Hope in the States

The role of the federal government in regulating artificial intelligence has been marked by both advances and setbacks. Under President Joe Biden, the Department of Labor issued recommendations in October 2024 urging companies to disclose their use of AI, involve employees in strategic decisions, and support workers in positions threatened by automation. These guidelines were designed to establish transparency and safeguard labor rights at a time when AI was rapidly expanding across industries.

However, in January 2025, just hours after returning to the White House, President Donald Trump revoked the executive order that served as the foundation for those recommendations.

With a single move, the framework was dismantled, leaving unions and worker advocacy groups concerned about the absence of federal protections. Trump’s presidential order sets out a new federal policy on artificial intelligence, emphasizing that the US must remain the global leader in AI by ensuring systems are developed free from “ideological bias and regulatory barriers.”

Despite this, there is movement for AI regulation at the state level. “There have been hundreds of bills introduced to try to responsibly regulate AI in the states,” said Lauren McFerran, executive director of the AFL-CIO Tech Institute, a nonprofit organization affiliated with the AFL‑CIO but operating independently as a 501c3 entity. The Tech Institute was created by the AFL‑CIO to serve as a center for the labor movement around technology and artificial intelligence.

McFerran explained that the most successful regulation efforts so far have been sectoral, targeted legislation designed to protect specific groups of workers. She pointed to California, where lawmakers recently sent Gov. Gavin Newsom a package of significant AI bills, including the No Robot Bosses Act on algorithmic management and AB-1331 addressing workplace surveillance.

McFerran noted that New Jersey has also taken action, passing a law to temporarily ban electronic shelf labels in retail stores—a technology that enables instantaneous price changes and also threatens tens of thousands of jobs and billions in wages. Similarly, she highlighted New York’s FAIR News Act, which requires labeling and disclosure of AI-generated journalism to protect both consumers and workers.

These state-level victories represent important steps in safeguarding labor against the risks posed by artificial intelligence.

Accountability

Even company managers are now echoing the concerns about AI that have spread throughout society. According to Dario Amodei, CEO of Anthropic and creator of the large language model (LLM) Claude, generative AI could lead to the disappearance of half of all entry-level, white-collar jobs. Amodei said that unemployment rates could rise by 10 to 20 percent if governments and companies fail to establish AI safeguards.

Anthropic recently said that it had blocked attempts to misuse artificial intelligence for the development of biological weapons. This disclosure underscores both the dual-use risks of AI, and the importance of building strong safeguards so that such technology cannot be weaponized. Allegedly, one of Anthropic’s scientists, Jacob Coxon, resigned and publicly criticized the industry for moving “too fast without responsibility,” a statement that quickly reverberated.

Senator Amy Klobuchar is leading efforts to require independent audits of AI systems before they are deployed. Republicans such as Ted Cruz and John Thune have joined the push, though the bill has most recently stalled due to the parties remaining far apart on key aspects. In California, Governor Newsom has already signed the state’s first law mandating independent audits of AI products.

These steps reflect growing concern that unchecked AI could lead to catastrophic outcomes. The recognition that AI can have dangerous consequences reinforces unions’ arguments that technology must be implemented with accountability, ensuring that innovation does not come at the expense of jobs, rights, or public safety.