AI in HR: what does the EU AI Act require?

Artificial intelligence has long since arrived in human resources: whether automated applicant pre-selection, AI-assisted employee assessments or recruitment chatbots, companies are increasingly turning to smart technologies. But the new EU AI Act (KI-VO) has suddenly placed these applications under particular scrutiny.

AI in HR practice: between efficiency and risk

Imagine your company uses AI to screen applications. Sounds efficient, doesn’t it? The algorithm automatically filters out unsuitable profiles and highlights suitable candidates. But what if it later emerges that the AI systematically disadvantages women or older applicants? AI discrimination is a genuine risk — and the new EU regulation puts such cases under particular scrutiny.

The same applies to AI-assisted employee assessments. Suppose software analyses employees’ communication behaviour and awards points for “efficiency” or “engagement”. What happens if the criteria are unfair or cannot be understood transparently? Black-box AI is a no-go when it comes to HR decisions.

Risk classification of AI systems in HR

The EU AI Act follows a risk-based approach and divides AI systems into different risk categories:

Unacceptable risk

AI systems deemed unacceptable are prohibited. These include systems that manipulate or exploit people by taking advantage of their vulnerabilities. AI-assisted social scoring systems that assess individuals on the basis of their behaviour or other factors and could discriminate against them are also prohibited.

High risk

AI systems used in sensitive areas such as HR are considered high-risk. Examples include systems for assessing applicants, automated decision-making on promotions or dismissals, and software for monitoring employees. These applications are subject to particularly strict transparency and control requirements.

Low to minimal risk

AI systems that merely perform supporting functions, such as automatic scheduling suggestions or chatbots for applicant enquiries, fall into the low- or minimal-risk category. They are subject to hardly any regulatory restrictions.

The EU AI Act: why HR AI is high-risk

The EU AI Act classifies AI systems in HR departments as “high-risk applications” because they have a direct impact on fundamental human rights such as access to the labour market, fair working conditions and the protection of privacy. Particularly in applicant selection, performance assessment and employee monitoring, there is a risk of discrimination, lack of transparency and inappropriate control. This means that companies using such technologies must meet strict requirements.

These include:

Transparency

The criteria on which the AI bases its decisions must be clear.

Non-discrimination

The AI must not disadvantage any groups.

Human oversight

Important decisions must not be made by AI alone.

Documentation & risk management

Companies must carefully assess and document their use of AI.

The penalties for infringements are substantial: depending on the severity of the case, fines of up to EUR 35 million or 7% of worldwide annual revenue may be imposed. This even exceeds the range of potential GDPR penalties.

The EU AI Act: why HR AI is high-risk

Find out about the background, objectives and core content of the AI Act, as well as the associated challenges, in our white paper “Die EU-KI-Verordnung: Ein Clarius-Guide”, and benefit from our recommendations for companies.

Die EU-KI-Verordnung: Ein Clarius-Guide

What companies should do now

The introduction of the EU AI Act is fundamentally changing the way HR departments work. AI is increasingly being used in areas such as recruitment, performance assessment and employee development to optimise processes and improve the basis for decisions. While this technology offers numerous advantages, it also presents challenges, particularly with regard to fairness, transparency and data protection. Companies now face the task of implementing the new regulatory requirements, reviewing their existing systems and preparing their employees to use AI responsibly.

If you use AI in HR or are planning to do so, you should take action now at the latest. Specifically, this means:

1. Review your existing AI systems:

   Do they meet the new requirements? If not, make adjustments. In future, HR departments will be required to regularly review the AI tools they use for compliance with the EU AI Act. High-risk systems may need to be certified before they can be used in practice. This could mean additional administrative work for companies, but ensures that AI systems operate fairly and reliably.

2. Commit to transparency and fairness:

   A “black box” is not a solution. In future, companies will have to disclose when AI is involved in decision-making processes and ensure that particular groups are not systematically disadvantaged. Using AI transparently and fairly can have a positive effect on employer branding. Companies that use sustainable, ethical and compliant AI solutions can position themselves as attractive employers.

3. Train your HR team:

   AI literacy will be essential in future. HR employees need to engage more closely with how AI technologies work and with their possibilities and risks. Companies should offer targeted further education and training to ensure that HR professionals make the right decisions when using AI and identify potential problems at an early stage.

4. Seek legal advice:

   The requirements of the EU AI Act are complex — an individual risk analysis helps avoid fines. Trade unions and works councils (Betriebsräte) are also likely to play a greater role in shaping AI-assisted HR processes. Companies should work with employee representatives at an early stage to ensure that the systems used comply with the law and do not disadvantage employees.

5. Adapt recruitment processes:

   Many companies already use AI in application processes, for example to analyse CVs or pre-select suitable candidates. The EU AI Act requires such systems to be made more transparent and understandable. This not only helps ensure compliance, but also improves potential applicants’ trust in the recruitment process.

The EU AI Act: both an opportunity and a challenge for HR

The EU AI Act places new requirements on the use of AI in HR. While the regulatory burden is increasing, the Act also offers an opportunity: companies that implement the EU AI Act proactively can position themselves as pioneers in the ethical and transparent use of AI. HR departments should familiarise themselves with the new requirements at an early stage in order to avoid compliance risks and strengthen applicants’ and employees’ trust in AI-assisted HR processes.

Ultimately, the future of HR will be shaped by a successful combination of human expertise and technological support — with the right balance between automation and human oversight. Companies that achieve this balance successfully can not only make efficiency gains, but also reliably meet compliance requirements.

However, the requirements of the EU AI Act are complex — an individual risk analysis with an implementation strategy tailored to your company helps avoid fines.

Benefit from the expertise of our compliance specialists!

So that you do not have to fight your way through endless regulations, we offer:

• E-learning on AI literacy:
  Learn what matters when using AI in HR.
• Legal advice:
  We review your AI applications and help implement the EU requirements.
• Risk assessment & compliance check:
  Make sure your HR AI is legally robust.
• Development of company guidelines on the use of AI
  Let us create clear parameters together

Let us work together to ensure that your AI remains efficient, fair and legally compliant. We will be glad to advise you.

Find out more