News

(AI) AI refers to systems that display intelligent behaviour by analysing their environment and taking actions – with some degree of autonomy – to achieve specific goals. AI-based systems can be purely software-based, acting in the virtual world (e.g. voice assistants, image analysis software, search engines, speech and face recognition systems) or AI can be embedded in hardware devices (e.g. advanced robots, autonomous cars, drones or Internet of things applications). ( AI refers to systems that display intelligent behaviour by analysing their environment and taking actions – with some degree of autonomy – to achieve specific goals. AI-based systems can be purely software-based, acting in the virtual world (e.g. voice assistants, image analysis software, search engines, speech and face recognition systems) or AI can be embedded in hardware devices (e.g. advanced robots, autonomous cars, drones or Internet of things applications). )-based worker management (AIWM) offers benefits for productivity and operational optimisation, but it also introduces significant risks to occupational safety and health (OSH), particularly of a psychosocial nature. To ensure the successful implementation of AIWM systems while safeguarding worker safety and health, it is crucial to focus on worker participation and proactive prevention strategies. Additionally, European Union (EU) regulations designed to protect workers’ OSH are essential to mitigate these risks.

Key points for prevention 
To minimise the negative impact of AIWM on OSH, a human-centred approach is crucial, ensuring that these systems are designed and used in a way that preserves workers’ autonomy and job control. The technology should complement human decision-making and not replace workers or managers but support their roles. Moreover, considering AIWM’s implications for OSH from the early stages, following a ‘prevention through design’ approach, is recommended.

Raising awareness among all actors about the risks associated with AIWM, particularly psychosocial ones, is also vital. These technologies can lead, for example, to increased work intensity and performance pressure, as well as reduced support from peers and managers. This impacts safety and health by exacerbating fatigue and stress levels, increasing the risk of accidents and musculoskeletal disorders.

Holistic and regular workplace risk assessments, as mandated by the OSH Framework Directive, are important for identifying AIWM-related risks. These assessments should be conducted and updated regularly, as AI refers to systems that display intelligent behaviour by analysing their environment and taking actions – with some degree of autonomy – to achieve specific goals. AI-based systems can be purely software-based, acting in the virtual world (e.g. voice assistants, image analysis software, search engines, speech and face recognition systems) or AI can be embedded in hardware devices (e.g. advanced robots, autonomous cars, drones or Internet of things applications). systems evolve, to ensure that emerging risks are effectively managed as technology advances.

Providing adequate training for workers on how AIWM systems function, including the types of data they collect, their intended purposes, access rights, decision-making processes, system interaction, feedback provision, etc., is critical for promoting transparency, building Trust can be defined as the attitude that an agent [automation technology, i.e. advanced robotics] will help achieve an individual’s goal in a situation characterised by uncertainty and vulnerability. and ensuring fairness.

Finally, engaging workers and/or their representatives in the design, implementation and assessment of AIWM systems is essential for minimising OSH risks. By involving them, organisations can ensure that AI refers to systems that display intelligent behaviour by analysing their environment and taking actions – with some degree of autonomy – to achieve specific goals. AI-based systems can be purely software-based, acting in the virtual world (e.g. voice assistants, image analysis software, search engines, speech and face recognition systems) or AI can be embedded in hardware devices (e.g. advanced robots, autonomous cars, drones or Internet of things applications). systems are developed with consideration for workers’ needs and concerns. This engagement helps prevent excessive surveillance and control and helps preserve workers’ autonomy, safety and health.

Regulations contributing to the prevention of OSH risks of AIWM
A number of existing regulations play a key role in protecting workers’ safety and health from the risks of AIWM systems.

• The OSH Framework Directive requires employers to ensure the safety and health of workers from all risks at work, which implicitly includes those related to the use of AIWM systems in the workplace. Employers are therefore obligated to assess all potential risks stemming from AIWM systems, including psychosocial ones such as high workload, work intensity and reduced autonomy and take preventive actions. Furthermore, the Directive mandates that employers consult with workers and provide necessary training to protect their safety and health, including in relation to AIWM systems and their OSH impacts.
• The EU Artificial Intelligence Act classifies as ‘high risk’ AI refers to systems that display intelligent behaviour by analysing their environment and taking actions – with some degree of autonomy – to achieve specific goals. AI-based systems can be purely software-based, acting in the virtual world (e.g. voice assistants, image analysis software, search engines, speech and face recognition systems) or AI can be embedded in hardware devices (e.g. advanced robots, autonomous cars, drones or Internet of things applications). systems used for decisions on promotions, terminations, task allocation, performance monitoring and worker behaviour assessment, some of these purposes being related to how AIWM systems are sometimes used. This designation imposes mandatory requirements, including maintaining risk management processes throughout the system’s lifecycle and ensuring human oversight during AI training to minimise adverse effects on workers.
• The General Data Protection Regulation (GDPR) safeguards workers’ privacy also in the context of AIWM by regulating the collection and processing of workers’ personal data. In addition, it includes measures to prevent informational abuse resulting from faulty algorithmic decisions. Article 22 specifically limits the possibility of being subject to solely automatic decisions that produce legal effects.
• The EU Platform Work Directive, whose scope is limited to digital platform workers, obliges platforms to inform workers and their representatives about the functioning of algorithms used to manage work-related decisions. The directive also requires platforms to ensure adequate human oversight of automated decisions and provides workers the right to receive human explanations for any decision affecting them. This directive also establishes requirements for platforms to evaluate and mitigate risks associated with automated monitoring and decision-making systems.
• At the national level, the Spanish guide on Algorithmic information in the workplace highlights the obligation for employers to inform all workers about the use of algorithms in decision-making processes. This includes transparency about how algorithms are used for hiring, task assignments, productivity monitoring and promotions, ensuring that workers’ rights are respected and that any algorithmic decisions do not unfairly impact them.

The EU’s regulatory framework contributes to a safer implementation of AIWM systems. Beyond compliance with legislation, actively involving workers and applying preventive measures is essential to identify and manage the potential risks of AIWM to workers’ safety and health.

• Take a look at all content related to our current priority area, worker management through AI.
• Explore our previous HWC article focusing on the opportunities and risks of worker management through AI for OSH.
• #EUhealthyworkplaces continues on LinkedIn, Facebook, Bluesky and X.