Adding automation to manual labor changes the nature of the job, as well as the types of injuries workers routinely suffer.
Rapid advancements in robotics are changing the face of the world’s warehouses, as dangerous and physically taxing tasks are being reassigned en masse from humans to machines. Automation and digitization are nothing new in the logistics sector, or any sector heavily reliant on manual labor. Bosses prize automation because it can bring up to two- to four-fold gains in productivity. But workers can also benefit from the putative improvements in safety that come from shifting dangerous tasks onto non-human shoulders.
At least, that’s the story employers such as Amazon have, and largely successfully, promoted to the public.
In a recent study, Brad N. Greenwood, Dean’s Distinguished Professor at the Costello College of Business at George Mason University, investigated this question: Does automation make warehouse jobs safer? His co-authors include Gordon Burtch of Boston University and Kiron Ravindran of IE University. Their findings reveal that the answer depends on how safety is defined.
The researchers distinguish between two types of injuries: severe and non-severe. Severe injuries include broken bones, traumatic falls, and other incidents that cause employees to miss work. Non-severe injuries include sprains, strains, and repetitive motion problems, often leading to reassignment or light-duty work, but not missing work.
The findings showed that robots do seem to reduce severe injuries. In robotic fulfillment centers (FC), tasks like heavy lifting and long walks are handled by machines, reducing workers’ exposure to physical hazards. The researchers found a meaningful drop in the number of severe injuries in these facilities.
However, the overall picture is not so clear. In the same robotic warehouses, the researchers observed a sharp increase in non-severe injuries, especially during high-demand periods such as Amazon Prime Day and the winter holidays. The robotic fulfillment centers experienced a 40% decrease in severe injuries but a 77% increase in non-severe injuries compared to traditional centers.
To better understand their results, the researchers also analyzed thousands of online posts from Amazon warehouse workers. “There was an immediate and obvious discrepancy in worker opinion, based on whether their fulfillment center was roboticized or not,” says Greenwood.
Humans working alongside robots described their daily experience as “not physically exhausting” and “better than working at a legacy FC”. However, they also reported being expected to meet much higher performance metrics than their counterparts in non-automated FCs — amounting to a two-to-three-times higher “pick rate” in some cases. The faster pace of the human/robot dance was accompanied by a far more repetitive work routine that induced burnout in some workers, while causing others to “zone out”.
This dual reality — robots reducing some injuries while exacerbating others — has serious implications. For employers, simply introducing automation is not enough. Without careful job design, task rotation, and realistic performance goals, the shift to robotics can create new health and safety risks.
“Companies have bottom-line reasons to take this issue seriously. Beyond simple issues of liability, there is a cost to the firm of workers being unable to perform their duties,” says Greenwood.
Traditional safety metrics often focus on injuries that result in lost workdays. But as the nature of work changes, this approach may miss more subtle forms of harm. Chronic, repetitive injuries may not lead to time off, but they still decrease worker well-being and performance.
Looking ahead, Greenwood and his colleagues plan to explore how these trends play out over longer timeframes and in other industries. As robots become more common in fields like manufacturing, retail, and healthcare, similar patterns may emerge. The researchers hope their findings will help inform both corporate and public policy, ensuring that the future of work is not only more efficient but also humane.
“That isn’t to deny that warehouse robotics benefits workers,” Greenwood explains. “But we need to think more carefully about how to use them, and what that means for the humans they work with.”