A recent academic paper explores how Amazon has used technology to suppress union efforts at its Bessemer, Alabama, warehouse.
This study, authored by Teke Wiggin from Northwestern University, is the first of its kind to look into the role of algorithms in undermining pro-union movements.
It draws on many worker interviews and requests for information from the National Labor Relations Board.
In the 2021 vote, workers at the Bessemer facility decided against joining the Retail, Wholesale and Department Store Union (RWDSU), with a tally of 1,798 against and 738 in favor among 5,867 eligible voters.
This outcome indicates the effectiveness of Amazon’s extensive surveillance and control systems.
Since that election, Amazon has been instructed to conduct two additional votes at the site. However, the company has shown remarkable skill in thwarting worker organization using both traditional and modern methods.
Wiggin outlines several tactics, including using algorithmic discipline on worker devices, rapidly changing quotas, bombarding mobile devices with messages, and monitoring social media.
Amazon took advantage of workplace technology to manage and control employees. These devices were used to send anti-union communications, ask questions to assess union support, and make management meetings more intimidating.
During sessions where anti-union ideas were presented, workers were monitored for their efficiency and discipline, constantly reminded of the surveillance they were under.
Wiggin describes a tactic called “algorithmic slack-cutting,” which was employed to make working conditions seem more bearable.
This involved easing the enforcement of quotas and rules and stopping automatic firing for not meeting quotas, thus reducing the pressure workers felt.
Additionally, Amazon improved the previously automated HR experience by increasing the number of HR representatives present during the election to address worker concerns.
However, any relaxation of rules during the election could be reversed later. While Amazon tried to ease some disciplinary measures to respond to worker demands, it simultaneously increased its digital harassment through personal devices.
The company used an app that workers relied on for scheduling and pay information to send anti-union messages that could not be ignored.
The last strategy Amazon implemented involved monitoring social media interactions. According to Wiggin, the stress caused by AI and surveillance led workers to form online groups to share concerns and find ways to evade Amazon’s digital monitoring.
The study revealed that Amazon monitored over 43 Facebook groups, many of which were private, along with various websites and subreddits.
The goal was to “capture” and classify posts that could be investigated, especially those mentioning worker complaints or planned protests.
Wiggin shares stories of workers who were fired for posting about their working conditions on social media. He also mentions an “ambassador program” launched in 2018, where workers were paid to defend Amazon against criticism about work conditions.
Participants were trained to respond to any negative comments from customers, influencers, or media, emphasizing that those who worked at Amazon loved their jobs.
The use of AI and technology was not limited to monitoring workers but extended to tracking unions as well.
A memo from 2020 described an AI tool called geoSPatial Operating Console (SPOC), which was designed to analyze labor organization across different locations. This tool could visualize various data sets related to unions and labor activism.
Wiggin argues that his research shows Amazon is not merely adjusting existing AI systems to hinder unionization; instead, it is transforming and weaponizing technology to suppress dissent.
He notes that while companies can create real-time maps of unionization risks, they are also converting their tools into more oppressive mechanisms that dehumanize workers.
As Jeff Bezos of Amazon aligns himself with Donald Trump, the push for a unionized Bessemer warehouse continues into its fifth year.
Following claims that Amazon improperly influenced the first vote, the National Labor Relations Board called for a second election.
Disputes over the results of the 2022 election have led to plans for a third vote, though this now seems unlikely due to recent political actions affecting the board.
In closing, Wiggin describes Amazon’s management style as “despotic,” emphasizing its coercive nature. He points out that this environment is created not only by overt control methods like algorithmic management but also by subtle tactics that rely on underlying oppressive conditions, such as a harsh labor market and strict disciplinary measures.
Posts You May Like