The 1999 cult classic movie Office Space depicts Peter’s miserable existence as a software programmer who lives in a cubicle. The dreaded sentence, “I’m going to require you to go ahead and come in tomorrow,” is something Peter tries to avoid hearing from his boss every Friday.
Nearly 25 years later, this moment is still widely shared online because it perfectly depicts unsettling aspects of the working environment: Peter’s sense of impotence, the faux sympathy his supervisor uses to deliver this instruction, and the constant pressure to produce more.
Pop culture is rife with representations of deplorable employers. Even a movie with that name exists. However, things might be about to worsen. What should be made of the algorithm managers, the new bosses who are moving into offices across all industries?
algorithm management’s growth
Media coverage of the possibility of workers being replaced by robots is common. However, automation is not just happening in the labor market. And so are managers. Software algorithms are increasingly taking on administrative duties, including selecting applicants for jobs, assigning tasks, assessing worker performance, and even selecting when to terminate staff.
As monitoring and surveillance technology advances, tasks will increasingly be transferred from human managers to machines. especially wearable technologies that can monitor worker motion.
From the perspective of an employer, there are many benefits to delegating managerial responsibilities to algorithms. By automating operations that take longer for humans to execute, algorithms reduce corporate costs.
According to the most recent annual data, Uber can manage 3.5 million drivers with its 22,800 staff.
Artificial intelligence systems can also find solutions to improve corporate structures. The surge pricing strategy used by Uber, which temporarily raises fares to entice drivers during busy periods, is only viable because an algorithm is able to take into account variations in customer demand in real-time.
The dangers
More attention is paid to some algorithm management issues than others. Algorithmic bias is conceivably the risk that politicians, scholars, and journalists talk the most.
An infamous example is the now-defunct CV rating system at Amazon. This software, which utilized a scale of one to five to grade applicant CVs, was abandoned because it regularly gave male features on a CV a higher rating than those considered to be more feminine.
But a number of other problems are related to the development of algorithm management.
The first is the issue with transparency. Traditional algorithms only produce preprogrammed outputs and are programmed to make judgments based on step-by-step instructions.
On the other hand, machine learning algorithms acquire the ability to decide for themselves after being exposed to a large amount of training data. This implies that as they grow, they get more complex, and even programmers find it difficult to understand how they work.
When the justification for a choice, such as whether to fire an employee, is not clear, a questionable moral arrangement is in play. Was the algorithm’s choice to fire the worker arbitrary, unfair, or corrupt?
If that were the true, its output would generally be seen as morally repugnant if not criminal. But how might an employee prove that the reasons for their termination were illegal?
By sheltering abuses of power from redress, algorithm management exacerbates the power disparity between employers and employees. Additionally, algorithms remove from the work connection a crucial human function. It’s what the late philosopher Jean-Jacques Rousseau referred to as our “innate repugnance to witnessing one’s fellow human suffer” and “natural sense of sympathy.”
There is 0% likelihood that algorithm bosses will be sympathetic, even though not all human managers are. In our case study of Amazon Flex couriers, we saw the frustration platform employees feel when human appeals are rejected by the algorithm. Algorithms created to maximize efficiency don’t worry about unexpected childcare needs. Because they are still learning the job, they have no patience for employees who move slowly. They do not attempt to reach an agreement that aids a worker who is dealing with an illness or handicap.
How can we help?
Researchers, labor unions, and software developers aiming to improve safe working conditions have already made the hazards faced by employees managed by algorithms a major point of concern. Politicians in the US are debating expanding workers’ digital rights. Regular impact analyses of how algorithms influence employees and giving workers a vote in how these technologies are employed are two other solutions.
The necessity to generate a profit is not a justification for tolerating employee pain, despite the fact that firms may find management algorithms to be extremely profitable.
Peter eventually mastered the art of controlling his employer and elevating the workplace. He achieved this by demonstrating his value in warm, engaging interactions with senior management. How would he have performed if his supervisor had been an algorithm is the question.
