We are workers and civil society organizations, and together, we urge you to protect workers and center their rights as Congress considers AI
The Honorable Charles E. Schumer
United States Senate
Washington, D.C. 20510
Dear Leader Schumer,
We are workers and civil society organizations, and together, we urge you to protect workers and center their rights as Congress considers the promise and perils of artificial intelligence. Last month, members of Congress led by Senator Markey and Representative Jayapal sent a letter to leading AI corporations demanding answers about the labor conditions of data workers, and the processes by which data workers contribute to training and deploying AI. The contributions of data workers, often invisible to the public, are critical to advancements in AI. The corporations failed to adequately answer the questions posed by members of Congress. We therefore urge you to consider how workers, across sectors, are already impacted by new technologies and respond to their demands.
Increasingly, employers are using artificial intelligence and related technologies in ways that undermine working conditions. Corporations like Amazon use constant, invasive workplace surveillance in order to maintain control over workers. In fact, Amazon is using surveillance to retaliate against workers and undermine their protected rights to speak out and take collective action. Similarly, employees at Google and Walmart allege their employers embedded browser extensions and listening devices to quash organizing activities.
In addition to union busting, these massive corporations use electronic surveillance and automated management of workers and enforce a dangerous pace of work, which has spurred a worker injury crisis. Amazon uses its package scanners not only to track packages, but also to measure the number of seconds between each scan made by a worker, ensuring they maintain quotas as high as scanning and moving one item every 11 seconds over long shifts. Workers must ignore safety precautions in order to keep up. This breakneck pace of the work can cause repetitive stress injuries. Similarly, FedEx workers must wear a heavy scanner strapped to the forearm, a surveillance tool that directly causes repetitive stress injuries as they move heavy boxes.
Due to employers’ lack of transparency regarding the use of these tools, it is difficult to collect reliable statistics on the prevalence of surveillance systems. But recent surveys suggest close to 80 percent of employers use monitoring software to track employee performance and activity and nearly half of employees who are being monitored say these practices have a negative effect on their mental health. They describe being under coercive surveillance with an eye toward finding people to fire. A UPS driver said, “It’s like you’re fighting for your job every day.” Amazon’s surveillance-based relationship with its employees has been described as one of “control, humiliation and unabating anxiety.” This “management by stress” can create a constant low-grade panic that seeps into a worker’s private time and even invades their sleep. Decades of research shows that job strain, which occurs when workers face high job demands but have little control over their work, can lead to a wide variety of harmful health consequences, including anxiety, depression, cardiovascular disease, and hypertension. Elevated stress further compromises the ability of workers to prevent injuries.
In some industries, employers use automated management systems to hire and fire workers. Such tools often serve to further marginalized groups of workers that have historically faced higher levels of surveillance and scrutiny. These systems have penalized workers for wearing a headscarf, having a Black-sounding name, and mentioning a women’s college. Some workers have been fired by automated systems, sometimes without cause or method of recourse. In the gig economy, workers are subject to algorithmic wage discrimination, described by law professor Veena Dubal as “the use of granular data to produce unpredictable, variable, and personalized hourly pay.” These consequential automated processes are so opaque that delivery workers in New York City call gig companies their “jefe fantasma,” or ghost boss.
In the AI industry itself, data workers are subject to low wages, wage theft, intense surveillance, retaliation against organizing, and arbitrary punishment with no method for recourse. Some MTurk workers report making below federal minimum wage, while they receive no health insurance or benefits. As much as a third of workers’ time is spent on uncompensated work and many are unpaid without explanation, while they are aggressively surveilled down to the keystroke. Amazon Mechanical Turk workers have banded together to improve conditions, but Amazon and the employers on the platform have yet to comply.
Further, corporations use this technology in ways that puts workers at risk of being replaced, fissured (or “gigified”), deskilled, or paid less. Digital labor platforms like Uber and Amazon’s Flex use automated management to actualize invisible corporate control of workers, despite asserting that the workers are independent businesses. As a result, platform workers are stripped of a host of labor rights and protections, and prone to wage theft, discrimination, and on-the-job injury. Other employers have used AI to replace helpline workers, ultimately giving callers suffering from disordered eating harmful advice.
In some cases, workers have been able to fight back and steer technology outcomes in a better direction, and these victories offer a roadmap for expanding worker rights over technology. The Writers Guild of America, in its contract negotiations with the Alliance of Motion Picture and Television Producers, bargained specifically to ensure AI would not replace steps in their writing work. Ultimately, the union won a deal including worker rights to credit and compensation, regardless of whether AI is used in part of their work. These protections are crucial, and should be codified into law.
Finally, the normalization of workplace surveillance creates an environment where workers know they are constantly being watched and where employers can more easily engage in anti-union surveillance, making it more risky to speak out and organize for better conditions. The General Counsel for the National Labor Relations Board has warned that such uses of surveillance may chill workers’ legally protected organizing rights. This is particularly concerning for low-wage workers, gig workers, or otherwise typically marginalized including Black and Latinx workers, for whom organizing for better conditions is more risky.
When employers are able to use new technologies in these ways, it not only impacts those particular workers–it undermines our entire economy. Management practices spread rapidly, and as corporations like Amazon grow to become the largest employers in the private sector, and wield extraordinary power from the concentration of data processing capacity, AI applications that hurt workers may spur a race to the bottom, lowering standards for all employers.
To guard against this dystopian future, Congress should develop a new generation of economic policies and labor rights to prevent corporations like Amazon from leveraging tech-driven worker exploitation into profit and outcompeting rivals by taking the low road. Establishing robust protections related to workplace technology and rebalancing power between workers and employers could reorient the economy and tech innovation toward more equitable and sustainable outcomes.
In order to do this in the face of rapidly developing technology, Congress must act to:
- Prioritize the health, safety, and wages of data workers that develop and train AI.
- Fortify workers’ right to organize and bargain around these issues, without being subject to employer surveillance and union-busting.
- Protect against predatory surveillance and automated management practices that can harm our physical and mental health, undermine wages, and increase precarity.
- Address ‘black box’ processes that have already proven to discriminate: hiring, firing, wage determination.
- Change public policies that currently incentivize job replacement, deskilling, and fissuring (or “gigify”).
- Guard against low-road business models that concentrate wealth and power away from working people into the hands of corporations.
- Adopt a framework that addresses the underlying lack of transparency, due process and fairness, and job security in our at-will employment system in the U.S.
- Incentivize innovation that enhances worker well-being and shared economic prosperity.
Thank you for your attention to this matter, and your continued leadership.
Amazon Employees for Climate Justice
AI Now Institute
Alphabet Workers Union-CWA
Center for Popular Democracy
Center on Privacy & Technology at Georgetown Law
Data & Society
Distributed AI Research Institute (DAIR)
Electronic Privacy Information Center (EPIC)
Families for Freedom
Fight for the Future
Friends of the Earth
Open Markets Institute
Surveillance Technology Oversight Project
The Technology Oversight Project
Workers’ Algorithm Observatory at Princeton University