Studies show that if algorithms measure every gesture, the promise of simplifying processes risks becoming psychological pressure.

Algorithmic management systems have entered the workplace with the promise of automating repetitive tasks and lightening part of the operational load. A transformation that, in effect, is already well underway: according to a representative survey conducted by theOECD Between June and August 2024, out of 6.047 companies with at least 20 employees, active in all economic sectors in France, Germany, Italy, Spain, Japan, and the United States, 74 percent used at least one software program to issue instructions or monitor workers' activities.

Among managers using these tools, 70 percent report greater availability of information, while some case studies cited in the report show how, in the Italian and French healthcare systems, algorithmic technologies have helped better organize workloads, reducing pressure, especially in emergency rooms. These benefits are made possible precisely by the collection and analysis of a growing amount of data, which, however, when it concerns both processes and the people who carry them out, pave the way for increasingly pervasive forms of control.

Software and sensors allow for the recording of time, speed, movements, communications, and even data relating to workers' health or fatigue. A survey of over 5.000 British employees conducted as part of the Pissarides Review found that 43,6 percent perceived increased surveillance at their workplace. This is a grey area where innovation—also fueled by regulatory gaps yet to be filled—ends up amplifying an already existing psychosocial risk: lo stress produced by the feeling of being constantly observed.

From final performance to daily conduct: how workplace supervision is changing

If we think about it, to monitor a worker today it is no longer necessary to observe him directly, because the work itself, once digitalized, to leave a continuous trace of oneself, according to a dynamic now familiar even in our private lives whenever we post something online. Access times, duration of activities, speed of execution, movements, and communications easily flow into a single profile, which reconstructs how the worker achieved the final result and shifts control from final performance to daily behavior.

The extent of the phenomenon emerges from the report OECD 2025 Algorithmic Management in the WorkplaceAccording to the survey, 88 percent of U.S. companies use software to monitor employee work time or activity periods, a figure that drops to 44 percent among companies surveyed in France, Germany, Italy, and Spain, and to 19 percent in Japan. Even more marked differences emerge in monitoring the speed of execution, practiced by 72 percent, 15 percent, and 7 percent of companies, respectively.

The scope of observation can also extend to the content and tone of conversations, phone calls and emails, the location of employees and even indicators relating to health or attention level, bringing to light a much more detailed knowledge of the worker than that needed to simply verify the performance of a task.

What happens if reward and sanction criteria depend on the evaluation power of algorithms?

The most significant leap in terms of data collection occurs when measurement helps to establish what should be considered productive, thus transforming the available information into rankings and decisions capable of directly impacting the worker's path. United States, always according to theAlgorithmic Management in the Workplace, 90 percent of the companies interviewed use software to automate at least one evaluation function; more specifically, 83 percent use it to reward positive results, 74 percent to set objectives, 67 percent to penalize performance deemed insufficient, and 50 percent to develop rankings that are also visible to other employees.

The system, therefore, builds the parameters to which the worker must progressively adapt, concentrating both the availability of data and the definition of the criteria for interpreting it in the hands of the company. This information and decision-making asymmetry, compounded by the opaqueness of algorithmic mechanisms, on the one hand strengthens the employer's power and, on the other, makes it more difficult to understand or challenge an evaluation.

The psychological cost of continuously monitored productivity

The same perception of being monitored, however, does not permeate the labor market uniformly, as it tends to strengthen precisely in those environments where every activity can be translated into data and inserted into a centrally governed organizational chain. This is demonstrated by the results of the National Survey of Workplace Technology, presented in the study Estimating the Prevalence of Automated Management and Surveillance Technologies at Work and Their Impact on Workers' Well-BeingThis is a representative survey of 1.273 US workers conducted in July 2024, according to which 68 percent of respondents reported at least one form of electronic monitoring, with increased exposure to the transportation sector and healthcare support professions (two areas in which time and travel are an integral part of the performance).

Even more marked is the gap linked to the company size, since 88 percent of those working in organizations with at least a thousand employees report being monitored, compared to 43 percent of those working in organizations with fewer than ten. In this way, complex organizations use surveillance as a governance tool and transform the need to coordinate thousands of people into the ability to constantly monitor their activities and behaviors.

The result can only coincide with a different way of experiencing work time: among those whose productivity is continuously monitored by companies, 46 percent say they work at an unhealthy or unsafe speed, compared to 15 percent of those who work without electronic controls, while 53 percent report anxiety, compared to 41 percent. While not demonstrating a causal relationship, the data therefore outlines a clear association between the intensity of surveillance and psychological distress.

The future of algorithmic work? It hinges on the right to know.

The problem that is bound to accompany the spread of these systems will be the growing gap between what technology knows about the worker and what the worker knows about technologyThis paradox is already visible in the OECD report cited above, according to which 69 percent of European managers believe these tools do not use individual data, even though the majority of companies adopt software for at least one monitoring activity.

This contradiction leads the authors themselves to hypothesize an underestimation or incomplete representation of data collection, and becomes even more significant when considering the leeway afforded to employees: among managers who report using data-driven software, 65 percent report that workers cannot opt ​​out of data collection, 22 percent that they cannot access the information, and 30 percent that they cannot request corrections. In the United States, a staggering 90 percent of managers say companies do not allow workers to opt out of data collection.

If the worker is unaware of the information that informs a decision and cannot verify or correct it, the algorithm ends up constructing a professional representation over which the worker progressively loses any possibility of influencing it. For this reason, the possibility of knowing and managing these tools will become increasingly decisive with their diffusion:

“As new technologies become more deeply embedded in professional life, it becomes essential to understand how our interaction with them impacts quality of life.”

explains to The Guardian Christopher Pissarides, Nobel Prize winner in Economics and co-founder of the Institute for the Future of Work.

Here are three insights that might interest you:

Will AI steal our jobs or transform professions and skills?
Drones, robots and AI are joining the front lines alongside rescuers
Italy launches portal making labor laws accessible