The AI Regulation is the world's first set of rules on artificial intelligence. The EU Parliament adopted it by a large majority on 13 March 2024, followed by the EU Member States in the Council. The AI Act, as the Regulation is also called, is an important step in the regulation of artificial intelligence, but there are also challenges and gaps.

Protection of fundamental rights
The AI Regulation aims to ensure a high level of protection of health, safety and fundamental rights. AI systems are classified according to risk categories and obligations are set according to the risk. Furthermore, the Regulation sets out requirements for manufacturers and operators regarding, inter alia, the use of data, risk management, human oversight, documentation and transparency of AI systems. Labour law or social policy aspects are not regulated, however. Some applications classified as particularly high-risk will be banned. What is extraordinary about the AI Regulation is that it aims to protect not only the health and safety of citizens, but also fundamental rights, democracy and the rule of law.

AI in the workplace
In the legislative process, a number of changes have been implemented in comparison to the European Commission's proposal, which contribute to better consideration of workers' rights and the protection of vulnerable persons as well as to greater transparency and information. This includes in particular the prohibition of AI systems for emotion recognition in the workplace and in educational institutions. The AI Regulation also sets strict requirements for the development and use of high-risk AI systems in the workplace, employee management and vocational training. These include, for example, recruitment systems as well as performance assessment and behavior monitoring in the workplace. In particular, employers (i.e. operators) must inform the workers concerned before commissioning or using a high-risk AI systems in the workplace. Operators must also ensure that workers and other persons involved in the operation and use of AI systems on their behalf have a sufficient level of AI competence. 

Challenges and gaps
The AI Regulation is a framework that has yet to be fleshed out through technical norms and standards. AK calls for the involvement of workers representatives in the standardisation and codification processes. Areas that are not classified as high-risk remain largely unregulated in the AI Regulation, posing potential protection gaps for workers. From AK's point of view, this is problematic with regard to the increase of AI-based (so-called algorithmic) management systems: Products are constantly coming onto the market that generate massive amounts of data or combine them from various applications that entail observation and control or automated allocation of work. This also raises fundamental rights issues. In view of the rapid development of new technologies, it is crucial to closely monitor the further development of the technical specifications and to ensure that AI is human-centred and designed from the perspective of the employees. AK therefore calls for clear rules on the use of AI, including for those areas that are not classified as high-risk. The European Trade Union Confederation calls for a dedicated EU directive on algorithmic systems in the workplace.

Implementation of the AI Regulation
The AI Regulation will have to be fully applied from 2026, but some rules will come into effect earlier. Prohibitions on certain AI systems, such as emotion recognition in the workplace, apply as early as six months after entry into force; some other important rules will have to be applied from mid-2025, such as the rules on general purpose AI. To implement the AI Regulation, the EU Commission has set up an Artificial Intelligence Office. AK also calls for well-resourced authorities at the national level. When using AI in the workplace, consultation and participation of employees as well as their representatives at company level and beyond must be ensured. Measures are also needed to train employees and raise awareness among all citizens about the rights and risks associated with AI and the involvement of workers' representatives in the design of new technologies at all levels.

Further information:
AK Wien: KI-Verordnung aus Sicht der Beschäftigten (German only)
AK Info: Die KI-Verordnung und ihre Folgen für die Arbeitswelt (German only)
AK Aktuell: Künstliche Intelligenz am Arbeitsplatz (German only)
AK EUROPA: Agreement on the world's first set of harmonised rules on artificial intelligence
ETUC: AI Act sets stage for Directive on AI systems in the workplace  
ETUI: Artificial intelligence, labour and society