"What are you looking at?", CC BY 2.0, by Jonas B
Tech accountability advocates and whistleblowers are raising alarms daily about the dangers posed by unfettered deployment of artificial intelligence (AI), while tech companies have enjoyed the current lack of regulation. As digital technologies including AI are deployed in a growing range of settings including the workplace, they pose a particular threat to workers’ rights.
Across industries, the rise of intrusive digital surveillance and automated decision systems—or “bossware”—is transforming work in ominous ways. These developments have exposed critical gaps in U.S. worker protections, with workers reporting increased injuries, more precarity, coercive work environments, routine wage theft, and other harms related to their employers’ use of digital management tools. Employers like Amazon have also reportedly used workplace surveillance systems and algorithmic management tools to aid in union busting.
In response, in recent years, several unions have won protections in their collective bargaining agreements that limit or prohibit employers’ use of digital tools to surveil workers or monitor their productivity. Further, a few jurisdictions have proposed or enacted policies designed to protect workers from harmful or unfair uses of AI in the workplace. The AFL-CIO recently established a task force of its state federation affiliates to coordinate and aggressively advocate for improved regulation of workplace uses of AI across states. Despite these efforts, employers and businesses have largely adopted digital monitoring and workplace automated management tools without meaningful negotiation with workers or robust regulation.
To respond to growing harms posed by these systems, policymakers must both update existing foundational workplace protections and regulate new technologies directly. A new National Employment Law Project report details the far-ranging harms to workers posed by bossware systems and offers a policy agenda to address them.
One good place to start is by reforming our at-will employment system.
Under at-will employment, which is the norm across the United States, workers can be – and frequently are – fired without advance notice for any reason or no reason at all. Employers have no obligation to disclose performance standards, show that they are being applied fairly, or provide feedback and a chance to address performance issues before discharging workers.
Introducing bossware into this at-will employment system further deepens the power imbalance between workers and employers. When employers have no obligation to be fair or transparent when firing workers, and actively monitor workers’ every action while measuring workers against opaque standards, you can see how intrusive technologies heighten a climate of fear and enable exploitation in the workplace.
"Amazon Prime" by Stock Catalog, CC BY-SA 3.0
Amazon warehouse workers recount how they are subject to constant monitoring and assessment under poorly disclosed and constantly changing standards that track their every movement, minute-by-minute, and cultivate an atmosphere of fear of not meeting unknown quotas. Office workers are now routinely monitored with software that tracks their keyboard, mouse, and head movements, grabs screenshots and photos of them, and evaluates, compensates, and disciplines them based on the results. The same is true for everyone from therapists to hospice workers. And call center workers are now being evaluated and disciplined by AI based on the tone and emotion in their voices.
What these forms of surveillance have in common is that workers often have little information about the standards they are being held to, no access to review and correct the data being collected, and no recourse when data inaccuracies result in discipline or termination. Workers may be racing against a clock they can’t see or trying to satisfy an algorithm that generates flawed outputs.
In this new ”normal,” policymakers interested in addressing the harms of workplace surveillance and automated systems must look beyond simply establishing privacy protections for workers akin to those that protect consumers. They must instead tackle the underlying lack of transparency, due process and fairness in the at-will employment relationship, all of which have been further intensified by new forms of surveillance.
“Just cause” employment protections, a framework that is already picking up steam nationwide, does exactly this. It guarantees workers a right to advance notice of terminations, a good reason before losing their jobs, and severance pay when they are terminated. Just cause laws also require that performance standards be disclosed to workers before they can be evaluated on their basis, and any materials and data used for disciplinary decisions must be fully disclosed.
In an encouraging sign, some cities and states are starting to rein in the explosion of intrusive technologies in the workplace with new labor protections that move in this direction.
Since 2021, five states – California, Washington, New York, Minnesota, and Oregon – have passed innovative legislation targeting conditions in the warehousing industry, and similar legislation has advanced in several other states. These laws include provisions that begin to establish disciplinary transparency guardrails. Some give workers the right to obtain comparative data from their workplaces to help ensure they are not being singled out unfairly.
These new warehouse worker disciplinary transparency protections are an important first step. But they should be expanded to all industries. And they need to be supplemented with full just cause protections against unfair discharge in order to make it realistic for workers to actually enforce these new guardrails.
Photo by Ecole polytechnique, CC BY-SA 2.0
These protections should be strengthened to include actual limits on the use of intrusive monitoring and technology. Certain types, such as video monitoring in a worker’s home and geofencing, should be categorically prohibited. For other types of monitoring, employers should be required to show they are reasonably necessary for evaluating performance. Productivity assessments should be measured only over reasonable increments of time – not the Orwellian, minute-by-minute monitoring of “time off task” that companies like Amazon currently use.
Labor advocates in a growing group of states are seeking to do just that. New legislative proposals like the Massachusetts’ Fostering Artificial Intelligence Responsibility (FAIR) Act and a package of California bills including the No Robot Bosses Act and the Workplace Surveillance Tools Act seek to prohibit some of the most abusive uses of surveillance and bossware, while providing for a more transparent process when it is deployed and requiring that important employment decisions such as discipline and discharge be made by humans, not robots.
But ultimately, to respond to the way surveillance and bossware are making jobs more precarious, workers need a fairer process governing discipline and discharge. To provide it, a small but growing movement led by worker organizations is pushing to rein in employment-at-will. New York City and Philadelphia have enacted “just cause” employment laws requiring employers in specific industries to provide a fair process, including advance notice, feedback, and a good reason, before workers lose their jobs. Now, New York City worker organizations and lawmakers are seeking to expand that law to workers in all industries – and to add new protections cracking down on abusive electronic monitoring in the workplace. Illinois worker organizations and lawmakers have introduced similar legislation.
Enacting just cause employment laws also could help to curb union busting. The at-will employment system, particularly when combined with weak enforcement of labor protections, makes it easy for employers to fire pro-union workers without explanation or accountability. Just cause laws requiring employers to provide a legitimate, documented reason for termination could make it far riskier for employers to disguise anti-union retaliation as routine discipline. Further, provisions that mandate transparency about performance standards and the data used in disciplinary decisions could expose when bossware data and outputs are selectively used to target union organizers or those sympathetic to unions. These safeguards would give workers a fairer chance to organize by lessening the fear of sudden, unjust dismissal. Just cause protections also reduce workplace turnover and churn -- factors that make it harder for workers to organize a workplace. Indeed, just cause laws are already protecting workers organizing on the job from unfair discipline. A New York City Starbucks worker who was active in organizing his workplace and subsequently unfairly fired won his job back under the City’s fast food just cause law.
In addition to adopting just cause laws, policymakers should also update and improve enforcement of existing laws governing worker health and safety, employment relationships, the right to organize, fair pay, and protections against abusive scheduling and overwork—to counter bossware-related harms.
Generative AI is spurring a long-overdue conversation about the damaging impact of uncontrolled technologies. Policymakers must heed workers’ demands and begin to reform our at-will employment system in order to rein in abuses of technology and prevent further degradation of job quality across industries.
