In commemoration of the annual World Day for Safety and Health at Work, a technical seminar was held on 28 April, organized by the National Institute for Occupational Safety and Health (INSST) in collaboration with the International Labour Organization (ILO), under the slogan “Revolution in the field of occupational safety and health.”[1] As is customary in this event, the presentation raised awareness of the various risks faced daily by workers and, in response to them, placed expectations on the crucial role to be played by artificial intelligence in reducing the number of work-related accidents. It was also emphasized that the fourth technological revolution, which defines our time, is inherently ambivalent: when used correctly, it can significantly enhance protection in digitalized employment by anticipating accidents through preventive data. Conversely, when misused, automated actions implemented without the necessary safeguards for fundamental rights at stake (e.g., privacy) would represent a clear regression, while also generating greater problems than those already existing.
Occupational risk prevention in the context of new technologies is not confined to platform work—where some regulatory progress has already been made—but, as is well known, digitalization exerts a transversal impact across all productive sectors. Preventive legislation, which has already been in force for thirty years (Law 31/1995, of 8 November, on the Prevention of Occupational Risks), should arguably be reviewed in order to address and respond to the new occupational risks stemming from AI—most notably psychosocial risks. Within this framework, it is undeniable that data protection and the reasonable handling of data by the employer must remain at the forefront.
Indeed, algorithmic management of workers is fundamental, as it enables the storage of vast amounts of data that are subsequently processed by AI. Such intelligence allows for automated decision-making which, without doubt, improves efficiency and productivity in enterprises. Examples of the utilities and opportunities offered by these digital technologies include, among others: improved scheduling and allocation of tasks, optimization of daily work, supervision of risks (particularly psychosocial risks), and data collection for problem identification and risk assessment. So much so, that the number of technological applications with tangible labour impacts is steadily increasing. To cite a practical example offering visibility: Enaible[2] measures how quickly workers complete assigned tasks and suggests ways to accelerate them; while Kronos AIMEE[3] predicts customer demand based on weather forecasts, providing recommendations on how many workers should be allocated per shift and which workers should be assigned to a specific task depending on their skills and capacities.
Jobs characterized by manual, routine, or repetitive tasks are the main target of algorithmic management systems (e.g., transport, warehousing, call centres, finance, etc.), and it is precisely over these that the principal risks associated with such technologies converge. The impact is both direct and steadily increasing in areas such as constant surveillance, reduced personal autonomy, heightened pressure in relation to performance evaluation and working time; but also in issues arising from lack of transparency, workers’ privacy, and data protection. These dynamics lead to greater demands, heavier workloads, and consequently, a higher probability of stress and mental health problems, which may become deeper and more complex.[4]
In order to ensure proper implementation of such systems, compliance with three key factors is essential:
- Consultation and participation of staff in all significant phases, in order to gain knowledge of the practical functioning of the technology.
- Transparency regarding data collection, its use, and the reasons behind it.
- Adoption of an appropriate occupational safety and health policy, together with a robust prevention management system.[5]
As can be seen, these three factors emphasize the human-centred approach, which not only ensures rational and respectful use of collected data but also raises awareness among all stakeholders, who are called upon to voice their concerns so as to correct and improve the technologies being employed. Yet caution is warranted: in the absence of compliance with these premises, as a consequence of this new organization of work, algorithmic management may become a breeding ground for legal disputes arising from work pressure. One may thus foresee an intensification of workload and its monitoring, together with a corresponding contraction of workers’ autonomy and trust, potentially generating situations of “techno-dependency,” in which the individual feels a high degree of insecurity because they are unable to carry out the task without the assistance of a machine.[6]
The challenges identified above are only some among many, with numerous others still unforeseeable today, which prevention professionals must remain alert to detect. What is currently known, however, is that preventive measures to reduce such psychosocial risks necessarily involve the integrated and effective management of such risks, together with the awareness-raising and consultation of workers, thereby ensuring their participation. In any case, a considerable number of applications are already in use which, provided data are managed appropriately, should serve to achieve accident prevention—the ultimate and overarching aim of this discipline.
References
