The use of artificial intelligence (AI) is already omnipresent in many areas of working life and in HR work. Nevertheless, German legislators have so far provided hardly any AI specific regulations in the context of employment. Employers in Germany are, however, not in a legal vacuum and must comply with various employment (and data protection) regulations when using AI. In future, employers will also have to observe the legal framework created by the recently adopted European Union’s AI Regulation laying down harmonised rules on artificial intelligence (AI Act).

AI-relevant employment regulations – Status quo

Here is a summary of the most important AI related employment laws which should be considered when implementing and using AI: 

General Equal Treatment Act (Allgemeines Gleichbehandlungsgesetz – AGG)

According to the AGG, discrimination against job applicants and employees during employment on the grounds of race or ethnic origin, gender, religion or belief, disability, age or sexual identity is prohibited (section 1, 7 AGG). Although AI is typically regarded as being objective, AI can also have “biases” and lead to discriminatory decisions. This may arise for example where the existing training data of the AI itself may be based on human biases. Currently, AI is, for example, often used to optimize application procedures. The use of a biased AI when selecting and rejecting applicants can lead to violations of the AGG which would be attributable to the employer.

In the event of a breach of the AGG, the employer is exposed to claims for damages and compensation payments. If the employee can establish facts which may suggest discrimination has occurred, the employer must prove that there was in fact no discrimination. This can be challenging where AI is involved.

Employers who use AI in a job application process or during employment should therefore ensure that the decisions made by the AI systems do not violate the statutory prohibitions of the AGG. It is critical to maintain a in human control of AI systems. This also applies from a data protection law perspective (see below).  

Works Constitution Act (Betriebsverfassungsgesetz – BetrVG)

In companies with works councils, the implementation and use of AI can trigger a number of co-determination rights that need to be observed. The Works Constitution Act (Betriebsverfassungsgesetz – BetrVG) already specifically mentions AI and we explain this further below.  

  • Section 90 para 1 No. 3 BetrVG requires the employer who plans to use AI in the company to inform the works council in good time and to provide  information.
  • Section 80 para 3 BetrVG stipulates that if the works council must assess the introduction or use of AI. Consulting an expert is considered necessary and the work council may seek assistance of an expert on the employer’s costs.   
  • The works council’s consent for use of AI in the context of personnel questionnaires and general assessment principles is required pursuant to section 94 BetrVG. For example, if AI-supported procedures are used in the context of recruitment, the content of a corresponding questionnaire requires the consent of the works council.
  • The works council’s consent for use of AI for selection guidelines is required, section 95 para 2a BetrVG, e.g., selection guidelines for recruitment or transfers.
  • With respect to AI, the works council’s right of co-determination related to the introduction and use of technical equipment suitable to monitor the behavior or performance of employees (see section 87 para 1 No. 6 BetrVG) is likely one of the most relevant works council’s rights.
  • Last but not least, the implementation of AI could constitute a change in operation (Betriebsänderung) within the meaning of the BetrVG if the statutory conditions are met. The employer must in this case consult with the works council about the effects of the AI on the employees and, if applicable, conclude a social plan.

Data protection regulations

There are also important data protection regulations that must be observed which should only be briefly touched upon here. When using AI, the prohibition of an automated individual decision-making pursuant to Article 22 para 1 GDPR is particularly important. According to this section, a data subject shall have the right not to be subject to a decision based solely on automated processing, which produces legal effects concerning him/her or similarly significantly affects him/her.

Outlook on AI Act

The AI Act adopted by the Council of the European Union on May 21, 2024 is considered the world’s first comprehensive law regulating AI. The Act applies to both providers and deployers of AI systems. Employers will usually be consider deployers in terms of the AI Act. The Act does not stipulate a minimum size for the company or operation to be applicable.

At the centrepiece of the AI Act is the classification of AI systems into different risk levels. Different compliance requirements are then linked to the different risk tiers.

The AI systems used in the HR area will regularly be classified as so-called high-risk systems in accordance with Article 6 et seq. of the AI Act. This applies, among others, to AI systems used in the course of recruitment, task assignment, performance evaluation, promotion and termination of employment. Employers using high-risk AI systems will face specific compliance requirements under the Act. They must for in particular take appropriate technical and organisational measures to ensure they only use such systems in accordance with the instructions for use accompanying the systems. Moreover, they shall assign human oversight to natural persons who have the necessary competence, training and authority, as well as the necessary support. Further, an obligation to ensure that the input data corresponds to the intended purpose of the high-risk AI and is sufficiently representative with regard to the intended use exists (see Article 26 AI Act).

There are also general obligations that apply regardless of a specific risk class e.g., in relation to transparency and information, including duties to inform employees who are affected by an AI system about the AI use.  

Violations of the AI Act can result in severe fines

The AI Act will shortly be published in the official journal of the EU and enter into force twenty days after its publication. Even though most of the provisions will only apply two years after entry into force or later, it is prudent for employers to prepare by carrying out audits on the AI in use and prepare for the AI Act on time. It is also to be expected that works councils will increasingly deal with AI systems and demand corresponding information and/or training in the light of the AI Act.

If you have specific questions in this regard, the Reed Smith employment law team as the case may be together with our data protection colleagues will be happy to assist.