Your boss may soon become an AI program

(ORDO NEWS) — The 1999 cult classic Office Space follows the dreary life of Peter, a software engineer who lives in an office. Every Friday, Peter tries to avoid his boss and the scary words, “I need you to come tomorrow.”

The scene is still popular on the internet almost 25 years later because it captures the unsettling aspects of labor relations – the helplessness Peter feels, the false empathy with which his boss delivers this directive, the endless demand for better performance.

Pop culture has no shortage of images of terrible bosses. There is even a movie with the same name. But the situation could get even worse.

What to do about the new bosses that are taking over jobs across industries: algorithmic managers?

The emergence of algorithmic control

The prospect of robots replacing workers is often covered in the media. But not only labor is automated. Managers too.

Increasingly, we see how software algorithms take over management functions such as selecting job applications, delegating work, evaluating work performance, and even making employee termination decisions.

As surveillance and control devices become more sophisticated, the number of tasks shifted from people to machines will only increase. In particular, wearable technologies that can track the movements of employees.

From the employer’s point of view, shifting the responsibilities of managers to algorithms can bring significant benefits.

Algorithms reduce business costs by automating tasks that take a human longer to complete. Uber, with its 22,800 employees, can monitor 3.5 million drivers, according to the latest data for the year.

AI systems can also find ways to optimize business organization. Uber’s pricing model (temporary price increases to attract drivers during busy periods) is only possible because the algorithm can process changes in passenger demand in real time.


Some algorithm management issues attract more attention than others. Perhaps the risk most discussed by journalists, researchers, and politicians is algorithmic bias.

An infamous example is Amazon’s broken CV ranking system. This program, which was used to rate job applicants’ resumes on a scale of one to five, was discontinued because it consistently ranked resumes with masculine characteristics higher than similar resumes that were considered more feminine.

But there are still several problems associated with the development of algorithmic control.

One of them is the problem of transparency. Classical algorithms are programmed to make decisions based on step-by-step instructions and produce only programmed results.

Machine learning algorithms, on the other hand, learn to make decisions on their own after receiving a large amount of training data. This means that as they develop, they become more and more complex, making their operation opaque even to programmers.

When the rationale for a decision such as firing an employee is not transparent, it is a morally dubious mechanism. Was the algorithm’s decision to terminate the employee biased, corrupt, or arbitrary?

If so, then its result will be considered morally illegitimate, if not illegal in most cases. But how can an employee prove that his dismissal was the result of illegal motives?

Algorithmic governance exacerbates power imbalances between employers and employees by shielding abuses of power from redress. In addition, algorithms exclude the most important human function from labor relations.

This is what the late philosopher Jean-Jacques Rousseau called “the natural feeling of pity” and “an inborn aversion to seeing the suffering of one’s neighbor.”

Even if not all human managers are compassionate, the probability that algorithmic managers will be compassionate is zero.

In our study of Amazon Flex couriers, we observed how desperate platform workers react to the inability of the algorithm to accept human calls.

Algorithms designed to maximize efficiency are indifferent to childcare emergencies. They can’t stand it when workers move slowly because they are still learning how to work. They do not negotiate to find a solution that will help an employee struggling with an illness or disability.

What we can do

The risks faced by algorithm-driven workers are already the focus of researchers, trade unions and software developers who are trying to promote good working conditions. American politicians are discussing the expansion of digital rights for workers.

Other solutions include regularly assessing the impact of algorithms on workers and giving workers a say in how these technologies are used.

While businesses may find management algorithms highly lucrative, the need to make a profit is not a reason to endure employee suffering.

Peter eventually learned to manage his boss and make work enjoyable. He did this by demonstrating his value in very personal meetings with senior management. The question is, how would he cope if his boss was an algorithm?


Contact us: [email protected]

Our Standards, Terms of Use: Standard Terms And Conditions.