The EU AI Act: What should employers be paying attention to right now

Littler’s 2025 European Employer Survey shows that many employers feel unprepared for the EU AI Act (EU Regulation no. 2024/1689 of 13 June 2024)—even though AI use in HR and the workplace continues to rise. Examples include systems that analyze job applications, automate talentmanagement and repetitive tasks, and prevent workplace accidents. 
 
AI systems in the EU must comply with the GDPR. Since 1 August 2024, the AI Act has also been in force on top of that. Its obligations are being introduced gradually, with most employmentrelated rules applying from August 2026.

A riskbased approach

The AI Act classifies AI-systems into four categories: 
• Unacceptable risk 
• High risk 
• Medium risk 
• Minimal risk 
 
Unacceptablerisk systems are already prohibited, such as emotionrecognition tools in the workplace. 
 
Most AI systems used in employment and HR—for example, recruitment tools that screen or rank candidates—fall under the highrisk category. These will need to meet additional safeguards from August 2026.

Employers using highrisk AI systems must:

• Take appropriate technical and organizational measures to ensure they use the AI-systems in accordance with the instructions for use, ensure that input data is relevant and sufficiently representative, keep logs, and so on. 
• Assign human oversight to natural persons who have the necessary competence, training and authority, as well as the necessary support. 
• Ensure transparency and inform employees and employee representatives in advance in accordance with local law 
• Contribute to ongoing risk and safety management 

Beyond compliance

These obligations do not stand on their own. Compliance with the AI Act affects not only the proper use of AI systems, but also for example how employers safeguard employee wellbeing and health and safety. As a result, employers are required to actively contribute to the prevention and mitigation of the risks arising from the use of AI systems in the workplace. This includes mental strain and stress resulting from more complex work or increased performance pressure due to uncertainty about job security.

What should employers do now?

Now is an important moment for employers to map all AI-systems they use and identify the corresponding applicable risk categories. If this has not yet been done, an AI policy can be developed to ensure adequate AI literacy and transparency. These efforts should also clarify which employee representative bodies within the company need to be informed where appropriate.