A.I. & Working People
- JFK
- Sep 13
- 4 min read
A.I. Is Redefining Working People. It’s Time to Unionize Our Data As Labor
As I prepare to head to Scandinavia to lead policy talks on how to use Europe's Data Act for Organized Labor, I can't but think about my fellow Americans. Artificial intelligence is rewriting the rules of labor. With every keystroke, scroll, or verbal prompt, humans are producing the fuel that powers modern AI systems: data. Whether workers realize it or not, their behavioral patterns, digital communications, productivity stats, and biometric inputs are now core inputs in automated systems that generate enormous value for companies—and often, that value is extracted without compensation.
It’s time we call this what it is: labor.
In the book Data Is Labor I argue that like the vehicles UAW Union Members produce, our digital exhaust—what we’ve casually called “data” for decades—is not just metadata; it’s labor output. And this is no longer theoretical. Today’s most advanced AI systems are being trained and fine-tuned on the lived experiences of workers: the Teamsters Union’s truck drivers, whose geolocation, braking patterns, idle time, and route decisions are now embedded in autonomous trucking algorithms; and SEIU’s service workers, whose voice interactions and resolution behaviors in call centers are increasingly used to train generative AI chatbots and automated customer service agents. In both cases, the AI systems replacing or surveilling workers were trained on those very workers’ labor—data-driven and uncompensated.
Why AI Forces This Reckoning
AI systems don’t create value in a vacuum. They are trained on human-generated data—from social media posts to call center transcripts to internal workplace chat logs. These systems then automate tasks that previously required human labor, but paradoxically, they are only as valuable as the data we feed them. That makes every worker a co-producer in the emerging AI economy.
In the age of AI, data isn’t just a byproduct of labor. Data is labor itself.
This reframing matters because it changes who gets to benefit from the AI revolution. Will it be only platform owners and AI developers? Or will the people whose data makes these tools possible get a share of the upside?
A Legal Strategy for Labor Unions
Unions now have the opportunity—and the obligation—to protect their members from becoming uncompensated data sources for AI systems:
Recognize data generation as compensable labor. From surveillance systems to productivity dashboards, the modern workplace increasingly relies on worker-generated data for business intelligence and automation.
Negotiate data rights into union contracts. This includes revenue-sharing mechanisms, transparency requirements, and limits on how employee data can be used—especially for algorithmic management or training AI models.
Pursue statutory recognition of data labor at the state and federal levels, positioning it alongside traditional labor protections.
AI makes this case stronger: it clearly converts human data into products, outputs, and profits. The more sophisticated the AI, the more critical the human labor in the form of data becomes.
The European Model: A Blueprint for the U.S.
In January 2024, the European Union’s Data Act took effect, with many key provisions coming into force by September 2025. It grants individuals and businesses the right to access, port, and profit from data they generate—particularly when that data is used commercially. It also defines a standard for “reasonable compensation” when companies rely on user or worker data.
Though the EU approach is framed as a digital market regulation, its principles align closely with labor rights. American unions should use this as both a precedent and a provocation: If the EU can build rules around fair compensation for data use, why can’t the U.S.?
What Employers Should Do Now
Rather than waiting for litigation or legislation, American companies can lead the way by treating data rights as a new frontier of corporate ethics. Here’s how:
Pilot voluntary compensation schemes for employee data used to train AI systems or inform business intelligence.
Ensure transparency on how employee data is collected, used, and monetized—including the use of productivity monitoring tools and algorithmic decision-making.
Negotiate data clauses with unions in good faith, recognizing that workers deserve to share in the value they help create.
Establish ethical AI governance boards with employee and union representation.
Educate employees on how their data contributes to automation and AI—and how they can assert their rights.
This Is a Labor Issue—And a Civil Rights Issue
The stakes are broader than just wage fairness. The unchecked use of human data in AI systems has deep implications for privacy, consent, bias, and civil liberties. When workers are used as invisible inputs in opaque AI systems that determine hiring, promotions, and surveillance protocols, their autonomy is eroded—and traditional labor rights are rendered toothless.
This is especially true for Black, Latino, immigrant, and working-class employees who already face disproportionate workplace surveillance and bias. Without intervention, AI will supercharge inequality under the guise of efficiency.
A Turning Point
American unions were born in the factories of the industrial revolution. Today, they are being reborn in the data centers and call logs of the AI era. The question now is whether the workers who generate the fuel for AI will be fairly paid—or quietly replaced.
Let’s be clear: artificial intelligence is not inevitable. Its design, deployment, and reward structures are still being shaped. We must ensure they are shaped by justice.
Your data is your labor. And in the age of AI, labor owes you a payment.
Comments