Computer in command

Policy Study

Cross-country workers’ survey on consequences of algorithmic management

Computer in command: Consequences of algorithmic management for workers

76% of workers within sectors like warehousing and telemarketing experience one or more forms of Algorithmic Management (AM) in the workplace. They report having less autonomy in their jobs, a greater workload, feeling more stressed, and greater concern among employees about their job security. 

These are the findings from the survey “Computer in command: Consequences of algorithmic management for workers” conducted by the Foundation for European Progressive Studies (FEPS) and Nordic think tanks among over 6,000 union members in Denmark, Sweden, Norway, and Finland. This study is the first of its kind to systematically examine the consequences of AM.

“Computer in command: Consequences of algorithmic management for workers” uncovers the adverse consequences AM for workers and how companies can avoid these effects. It shows that this use of AM has several negative consequences for the employees, including lower levels of trust between employees and management, lower motivation and job satisfaction, and greater concern among employees about their job security.

“The role of the manager has transformed with AM,” says Magnus Thorn Jensen, senior analyst at the Danish think tank Cevea and the study’s lead author. “Managers have always had the ability to monitor and control their employees, but new technologies have significantly expanded these possibilities. Our study shows that this can be detrimental to employee well-being.”

The survey indicates that some negative consequences for employees can be avoided if companies ensure a high degree of employee influence and transparency in management decisions. Where employee influence and transparency in management decisions are high, AM can be used without compromising employee job satisfaction, sense of job autonomy, and trust between management and employees. 

Whilst this is necessary, it alone is insufficient to ensure employee wellbeing. “There are limits to how intensive the surveillance and control that employees can be subjected to without harming the work environment,” says Asbjørn Sonne Nørgaard, director at Cevea and co-author of the report. “Therefore, there is a need to look at regulating how new technology is used in the workplace – both in collective agreements, legislation, and within individual companies.”

Whilst AM originated in the platform sector, the phenomenon has now spread across traditional sectors of the European labour market. The Platform Work Directive has set a precedent for a potential new European legislation on AI at the workplace, says Gerard Rinse Oosterwijk, digital policy analyst at FEPS and co-author of the report. “This study underlines the need for social dialogue on introducing and applying algorithmic management systems to mitigate some of the adverse effects on trust, autonomy, and job motivation, but also makes a case for legal guardrails and a potential new European directive to prevent adverse occupational health and safety risks of AM.”

Case studies for each country analysed in “Computer in command” are available below. Each study is focuses on a country based on quantitative surveys and focus groups provided in the language of the respective country.

Country case study – Survey: Denmark 🇩🇰

Computerchefer: Algoritmeledelse har store konsekvenser for medarbejderne

This publication focuses on the results of the survey among trade union members in Denmark. The study shows that Algorithmic Management (AM) is already quite widespread within warehouse work, customer service, telemarketing, and citizen services. A significant proportion of survey participants indicate that computer systems are used to assign shifts and tasks, monitor their activities, and evaluate their performance at work. This use of AM has a myriad of negative consequences for employees. They experience less autonomy in the performance of their daily tasks with an increased workload and stress level. Furthermore, this study shows that AM is associated with a diminished level of trust between employers and employees, a lower level of motivation and satisfaction in the job, as well as an increased fear of being fired for not reaching the expectations measured by the algorithm. Fortunately, these negative consequences are not entirely inevitable. The negative effects of AM are considerably less wherever there is a higher degree of employee influence in the decision-making and greater transparency in the management’s decisions. This study demonstrates that it is crucial for policy-makers, trade unions, and everyone in society to ensure that the use of new technologies in the labour market does not undermine employees’ well-being.

Country case study – Survey: Sweden 🇸🇪

Algoritmen som chef konsekvenser av algoritmiskt styrt arbete

This publication analyses the consequences of algorithmic management among Swedish warehouse workers. As the first large-scale quantitative analysis of its kind, the results of this study indicate that, on average, the use of AM has negative consequences for employees. The more workers are exposed to algorithmic control, the less autonomy they feel they have in their work, the less trust they feel from and towards their employer, and the greater workload they experience. Its use also affects employees’ well-being at work: When algorithmic control tools are widely used, employees are less satisfied with their jobs and less motivated, and they feel significantly more stressed and more uncertain about whether they will be able to keep their jobs.

But this study also shows that some of these consequences can be prevented through worker influence and transparency in the company’s decision-making processes. Therefore, it is of the utmost importance to ensure that the trade unions have sufficient tools to defend their members’ interests when it comes to the implementation of new digital systems in the workplace, both to protect the employees against a dystopian work environment where every aspect of the work can be monitored and controlled and to ensure that the Swedish model, with its large degree of trust between the parties in the labour market, stands strong in tomorrow’s digital reality.

Country case study – Survey: Finland 🇫🇮

Johtajana Tietokone: Algoritmisen johtamisen vaikutuksia työntekijöihin

In the report, we design a new algorithm management index based on survey data, which shows that algorithmic systems have a more widespread use for management in Finland than in other Nordic countries. On the other hand, according to the research, the negative effects of algorithmic management can also be prevented. Algorithmic management does not necessarily have to weaken the position or well-being of employees, but it requires that they are made aware of how it is used. Unfortunately, so far, that condition has not materialised well enough in the Finnish case, as less than half of respondents were satisfied with their employer’s communication or felt they could influence decisions about the use of algorithms.

–

Country case study – Survey: Norway 🇳🇴

Når sjefen bruker kunstig intelligens: Hvilke konsekvenser algoritmestyrt ledelse har for ansatte
By Hilde Nagell

This report, resulting from a survey of workers from the Norwegian context, shows that the use of algorithmic management can be associated to a number of negative consequences such as reduced job autonomy, increased workload, reduced job security, lower trust, reduced job satisfaction and motivation, and higher stress levels. Fortunately, such consequences are not inevitable; Workplace conditions appear to be particularly important in determining whether employees experience negative consequences.

In organizations where employees are involved and have influence in decision-making on the work floor, negative consequences are reported to a lesser extent. On the other hand, employees who report having significant influence on important decisions experience negative consequences from algorithmic management to a much lesser extent.

Secondly, the relationship between employees and managers appears to matter. Where there is a high level of transparency in management decisions, the negative consequences are also far less prominent, and in some cases completely absent.

Online appendix

Below are the research data and questionnaires employed in each targeted country.

–

This policy study series is part of the Digital Research Programme

FEPS, Tankesmedjan Tiden, Kalevi Sorsa Saatio, Tankesmien Agenda, CEVEA, Arbejderbevægelsens Erhvervsråd (ECLM), Friedrich-Ebert-Stiftung Nordics and Cooperation Committee of the Nordic Labour Movement (SAMAK), with the support of Nordic trade Unions, came together for a Digital Research Programme to discuss the research questions, methods and approaches of the different research strands:

1. Algorithmic management and governance (prevalence & consequences)
2. Workers’ experience at the workplace – surveys and focus groups
3. Online platforms and employment terms