Source: French to English Tester Published on: 2026-04-01
Source: The Conversation – France (in French)– By Dina Attia, Scientist, Senior Project Manager, French Agency for Food, Environmental and Occupational Health & Safety (Anses)
While the professional status of delivery workers is subject to increased scrutiny, the effects on their health of the algorithmic management implemented by platforms remain too little considered. Yet they are harmful, as highlighted by the National Agency for Health Security in its report dedicated to the issue.
Whether it is booking a rental for the weekend, finding a means of transportation in the middle of the night, or placing an order online, in recent years digital platforms have profoundly transformed our daily lives.
Meal delivery services, in particular,have experienced a rapid boom since their emergence in the early 2010s. The platforms that offer them have attracted thousands of workers and changed the way consumers access food services.
For a long time, debates about these transformations have mainly focused on the competition generated by these new market entrants or on the professional status of the delivery workers. The working conditions of the latter and their effects on their health have until recently been relegated to the background.
The conclusions ofthe Health-Competition investigation, published at the end of March 2026, shed stark light onthe deplorable working conditions of delivery workers(63 weekly hours, six to seven days a week, for an income well below the poverty line). To complement these quantitative data, it is useful to revisit the conclusions ofthe expertise that the National Agency for Food, Environmental and Occupational Health Safety (Anses) had devoted to this issue in March 2025.
The agency had indeed described the health effects of these difficult working conditions, and had deciphered the underlying mechanisms. A cornerstone of the activity of these platforms, thealgorithmic management(in other words, in this context, the use of computer algorithms to manage the activity of delivery workers) was particularly singled out due to its impact on health.
This system not only generates psychosocial risks, but it also exacerbates other harms to workers, concerning their physical health and social relationships.
Algorithmic management: a source of chronic stress
According to the European Union agency for occupational safety and health (EU-OSHA), thealgorithmic managementcan be defined as “the use of algorithms to assign, monitor, and evaluate work tasks as well as to monitor and assess worker behavior and performance […]”.
This organization, established by all the delivery platforms, creates an information asymmetry between the platforms and the couriers. Because the algorithm’s operation is governed by opaque and changing rules, the couriers have little or no possibility of appeal or human dialogue.
The work by Anses reveals that, in fact, this mode of management subjects workers to constant pressure.
The incessant notifications, the automated evaluations based notably on customer feedback, which do not take into account the realities on the ground, and the fear of account deactivation create a state of hypervigilance.
For illustration purposes, the delivery person cannot verify the packaging conditions of goods packed in a bag sealed by the restaurateur. However, the contents can spill during transport, causing customer dissatisfaction. Added to this is the fact that delivery people do not have clear information on how the ratings and comments left by customers after each delivery (or possibly other aspects of their service) will affect any potential sanctions.
The inability to anticipate income or said sanctions generates permanent feelings of powerlessness and economic insecurity. To maximize their earnings, delivery workers therefore adopt risky strategies based on “self-acceleration.”
In other words, they choose to reduce their break times or increase their work pace, hoping to be better evaluated by the algorithms. A situation that promotes professional exhaustion and the emergence of effects on mental health (chronic anxiety states, risk of burnout).
Other health effects: an alarming picture
Beyond mental health, Anses’s expertise highlights other health risks related to delivery activity. Thus, this expertise identified that26.4% of delivery workers in Île-de-France have been involved in a road accident.
Their delivery activity is also responsible for musculoskeletal disorders, such as lower back pain and tendinopathies.
It also disrupts theircircadian rhythmsCircadian rhythmsare cycles with a duration close to 24 hours that regulate many biological processes: alternation of wakefulness/sleep, variations in body temperature, blood pressure, hormone production, heart rate, and influence memory, mood, cognitive abilities, etc., which can result in chronic fatigue and, in the longer term, deleterious metabolic effects)
Finally, delivery workers are particularly exposed to urban pollution, which results in increased respiratory and cardiovascular risks.
These health effects are further exacerbated by the legal status of the delivery workers: by imposing the status of “independent” on them, the platforms transfer almost all of the work-related responsibilities to them.
The socio-legal determinants of risk
The platforms have limited involvement regarding prevention and remediation related to workplace accidents. More generally, there is no monitoring of the health of delivery workers.
Furthermore, this structuring of work limits the possibility of building collectives or mutual aid strategies, which further increases the professional isolation of workers.
Such conditions have repercussions on the physical and mental health as well as on the personal and family life of the delivery workers.
The situation is even more critical for some of the delivery workers, namely those who work “under a rented account.” These individuals, who cannot create their own account, rent the identity and the account of a delivery person officially registered on the platform, in exchange for a commission, because they cannot create their own account. According to the figures from the Health-Delivery survey, this situation affects about three quarters of the delivery workers,would cost them an average of 528 euros per month, on a gross income of 1,480 euros.
Often undocumented, these workers combine a total lack of social protection, an economic overdependence on the account renter, statistical invisibility (no data concerning them), increased risk-taking (more hours, fewer breaks), and an inability to assert rights or report an accident. This situation represents a major public health issue and a blind spot in current prevention mechanisms.
To improve the situation of delivery workers, Anses emphasizes the need for mandatory affiliation to a professional social security scheme, regardless of status (employee or independent). It also recommends coverage of occupational accidents for independent workers and the provision of rest and meeting places to facilitate mutual aid.
Experts also highlight the risk that AI tools pose to the world of work.
What regulation for AI in the professional environment?
The AI tools used by digital platforms now make it possible to measure a certain form of productivity in real time.
In logistics, for example, algorithms analyze every movement of order pickers in warehouses; in call centers, AI evaluates the tone of voice, speech rate, and conversion rate of telemarketers; in some offices, software measures keystrokes, mouse clicks, or the time spent on each application..
This quantification raises questions because it reduces work to measurable metrics but ignores the relational, creative, or reflective dimensions of professional activity. Good customer service is not limited to a short call time, and the quality of intellectual work cannot be measured using sometimes simplistic indicators…
Faced with these abuses, the European legal framework is evolving. TheAI Act (EU Regulation 2024/1689), progressively effective since 2024, classifies AI systems used to evaluate employee performance among the “high-risk systems” requiring strict regulation.
In France, starting from August 2, 2026, any worker subject to an evaluation by artificial intelligence must be informed of it. AI will no longer be able to operate completely autonomously: active human supervision will be mandatory, and no decision affecting an employee (promotion, sanction, dismissal) can be based exclusively on an algorithm.
Risks that concern all workers
The pressure of constant performance, without the possibility of disconnecting, creates a climate of permanent anxiety and exposes workers to chronic stress. As shown by thestudies on algorithmic management, automated decisions (task assignments, evaluations)reduce the workers’ room for maneuverand lead to a loss of autonomy.
The use of artificial intelligence also facilitates the fragmentation of contracts (micro-tasks, hybrid statuses), limiting access to social protections and leading to the precarization of workers’ statuses.
Finally, the absence of human interactions and work groups exacerbates loneliness and the feeling of insecurity, resulting in their isolation and a dehumanization of the professional environment.
While platform workers are currently the main ones affected by this situation, these practices could soon extend to many sectors of activity and types of employment.
What future for work?
Beyond economic or technological issues, algorithmic management and the working conditions implemented by delivery platforms reflect a societal challenge. As we have seen, the platform model, based on flexibility and hyper-optimization, externalizes risks onto the delivery workers. A pressing question arises today: how to reconcile innovation and the protection of workers in an increasingly digitized world?
While theEuropean Directive 2024/2831 on improving working conditions in the framework of workviaa platformmust be transposed into national law by November 2026, public actors have a window of opportunity to integrate these health issues into national texts and encourage the accountability of platforms.
Without a regulatory framework, the implementation of AI at work risks extending these abuses to other workers, in all sectors. With the increasing introduction of AI in companies, employees are increasingly exposed to mechanisms of algorithmic surveillance, automated scoring, and precarious statuses (short-term contracts, independents). Intrusive surveillance, constant pressure, and lack of recourse could become widespread.
The issue is therefore not only to protect platform workers, but to rethink work in the digital age in order to prevent the emergence of a society where surveillance, precariousness, and stress become the norm. It is the future of work that is at stake.
The authors do not work for, advise, hold shares in, or receive funds from any organization that could benefit from this article, and have declared no other affiliation than their research institution.
–ref. Home delivery workers: how “algorithmic management” deteriorates workers’ health –https://theconversation.com/home-delivery-workers-how-algorithmic-management-degrades-workers-health-279469