How will new advanced information technologies impact work? This is a major focus of public debate right now, driven by widespread fears that automation will soon leave tens of millions unemployed. But debate so far has tended to neglect the relationship among technological innovation, political economy, and the law of work. This is a major omission, since the automation of particular tasks doesn’t just happen. Rather, it takes place under laws that are subject to democratic oversight and revision – and with different laws, we could encourage a radically different path of technological development, one in which workers have a real voice, and in which they share consistently in technology-driven productivity gains.
Take two upcoming transformations that we’re all familiar with: the automation of some kinds of driving and some kinds of fast-food work. Within a few years, truckers, delivery drivers, and taxi drivers may be able to use an autonomous mode consistently on highways. Later on, they may be able to do so on major suburban and rural streets. But given the wide variation in road quality, humans will likely need to pilot vehicles in residential areas and on city streets for some time to come. And given the wide variation in building structures that delivery robots would need to navigate, humans will almost certainly need to complete deliveries in many instances.
Similarly, in fast food, ordering kiosks are already displacing cashiers, but not in their entirety. Some customers are unable to use the kiosks, including the 70% of McDonalds customers who use the drive-through. Sometimes the kiosks will break down, and sometimes orders won’t be processed appropriately, and thus workers will need to step in. Food preparation may also be automated in part, but given the fine motor control and tacit knowledge required for cooking, it has proven resistant to full automation. Like the transformation in driving, then, this change will likely be gradual and iterative. Technology will augment human capabilities rather than replacing humans wholesale, and workers, companies, and consumers will need to adapt over time.
This process of technological change implicates many different legal and social regimes, including those related to workplace privacy and racial and gender equality, which I hope to take up in future posts. For now, I’ll focus on how labor and employment law allocate decision-making rights around the introduction of new technologies. The doctrinal answer is simple: in the U.S., firms have near-plenary power to reorganize production processes without input from workers. That allocation of entitlements has a powerful effect on workers’ welfare, and on other important social goods.
Two aspects of American labor and employment law play an especially important role here. The first is the American rule of “employment-at-will.” Absent a clear agreement to the contrary, either party to an employment contract can terminate it, at any time, for any reason at all, or even for no reason or a malicious reason, so long as doing so is not otherwise unlawful. Employment-at-will is no longer as robust as it once was, due to developments in the common law and due to norms against unreasonable termination, but it is still the critical background rule. Because of that rule, employers are free in many instances to walk in one day and offer workers a take-it-or-leave-it-deal: sign an agreement to arbitrate work-related disputes, or be terminated. The same logic transfers to the automation context. Accept this new technology, or find yourself a new job.
The second is a deeply rooted assumption that management “owns” the enterprise in a sense familiar from classical legal thought, or even from feudalism. This assumption is most apparent in labor law — the law of collective bargaining — but shapes the non-union workplace even more powerfully. As a result, decisions about the overall direction of the business such as investment strategy, product lines, production planning, and even whether to shut it down simply to avoid unionization become matters of unilateral managerial discretion, not subject to challenge by workers’ organizations, certainly not subject to challenge by workers themselves, and not reviewable by agencies or courts of law. The introduction of labor-saving technologies clearly falls within that zone. To be clear, under U.S. labor law, unionized workers do have rights to bargain over the effects of such technologies—for example, over which workers will be laid off or reassigned. Unionized workers also have clear rights to bargain over technologically-driven changes to terms and conditions of employment such as wages, hours, and scheduling practices. But since so few workers are unionized, those rights have little impact on the ground today.
These doctrines are so well established as to be unquestioned in public debates, but they reflect a contingent allocation of entitlements, one that many other advanced industrial nations have rejected. Elsewhere, workers often enjoy “just cause” termination protections, and therefore have greater space to contest the introduction of new technologies. Likewise, many European workers enjoy consultative rights about aspects of business management regardless of their unionization status, which again gives them a voice in how new technologies are deployed. In some countries, workers are even guaranteed seats on company boards, and therefore a voice in the broader arc of technological development.
The usual objection to such rules, of course, is that granting workers rights to contest technological innovation will thwart progress. But that need not be the case. As noted above, U.S. labor law reflects a sensible first cut at this issue: unionized workers can bargain over the effects of technological change but not over that change itself. Making that right universal would advance three distinct goals.
First, worker voice can in some instances enhance rather than hinder productivity. Unionized workers can complain about harmful or irrational workplace practices to management without fearing retaliation, which in many instances can ensure a more harmonious workplace and greater productivity. Indeed, the voice of workers may be especially beneficial in modern high-tech workplaces, where their input concerning workflow and processes can substantially enhance productivity. This logic overlaps with theories of user-driven innovation, under which users often adapt new products to their own idiosyncratic needs, then often feed their own innovations back to manufacturers. For example, giving drivers the power to let their employers or principals know about emerging problems with autonomous driving technologies could help ensure that those technologies are being used safely and to their maximum potential.
Worker voice can also enhance distributive fairness. Where new technologies enhance labor productivity, better bargaining rights should enable workers to capture a greater portion of profits, mitigating economic inequalities. For example, as driving tasks are automated, unionized drivers could push for severance packages, or higher wages, particularly if task automation enables companies to keep vehicles on the road for much longer periods of the day. In fast food, we could see workers pushing for a significantly higher wage floor as ordering kiosks are put into use. After all, those kiosks should save costs, and those savings could be passed on to workers rather than consumers and investors. In the many sectors that may be impacted by self-driving cars, we could again see workers demanding higher wages, fair severance pay, and a humane pace of work.
A similar logic applies to scheduling practices. The introduction of algorithmic scheduling by firms such as Starbucks has led to frequent worker complaints about irrational shifts – for example, working three hours, having two off, then working another eight; or being required to clock out during times of low customer demand. Unionized workers might push for longer guaranteed hours, or higher pay with scheduling flexibility, or some other preferred arrangement. But non-unionized workers are essentially at their employers’ mercy in these matters. Giving workers a voice can therefore ensure that new technologies are used to ensure productivity gains for all rather than merely being used to sweat workers.
Finally, worker voice can enhance democratic legitimacy. A national union of drivers could help develop national policy for displaced workers, retraining programs, or even worker-owned cooperatives that would manage fleets of semi-automated vehicles. A national union of fast food and restaurant workers could play a similar role with automated ordering and production systems. Such efforts would replicate at the macro-level some of the micro-level benefits of worker voice within individual firms. To be clear, if task automation really does begin to harm workers, the optimal policy approach may be to slow the pace of technological change so that they can adapt. But even in that case, worker voice will be critical to helping policymakers strike the right balance.
Making worker voice universal is a tall order in the U.S., given our dysfunctional labor law regime, but the necessary reforms are clear enough. We need to make it far easier for workers to obtain union representation, or even simply make representation a matter of right. We also need to make it easier for workers to bargain at the multi-firm or even the national level, so that unionized firms aren’t put at a competitive disadvantage. These ideas are politically utopian at the moment, but they are gaining traction among policymakers today. Again, any such reforms will interact in complex ways with our racialized and gendered labor markets, a topic I hope to consider in detail in a future post.
In one of the first posts on this blog, Purdy and Grewal criticized the neoliberal trope that “democratic politics and public institutions” simply cannot “successfully discipline and shape economic life.” In addition to regulating economic activity in the public interest, it is essential to incorporate democratic norms into day-to-day economic activity, especially in moments of dramatic technological and social change. In fact, it is hard to imagine an egalitarian and inclusive society without a robustly democratic economy, which requires giving workers more control over the technologies of their workplaces and beyond.