On May 12, 2026, another event in the “Automated Systems” lecture series took place in a hybrid format. The event was hosted by the Chair of Civil Law, European and International Business Law (Prof. Dr. Buck-Heeb) in collaboration with the Interdisciplinary Institute for Automated Systems (RifaS).
This time, the event featured attorney Marieke Mattern. Ms. Mattern has been an attorney at the law firm Dr. Krafczyk und Partner in Hanover since 2024. Since her admission to the bar in early 2022, she has advised national and international clients on all matters related to intellectual property law, IT law, and data protection. Before joining Dr. Krafczyk und Partner, she worked in the Digital Law and Litigation & Dispute Resolution practice at a Big Four firm in Hanover.
The presentation provided an overview of the key aspects that must be considered from a legal perspective in practice when drafting contracts regarding the use of artificial intelligence. Key topics covered in the presentation included the granting of usage rights, liability for damages and defects, and the protection of data and trade secrets.
When granting copyright usage rights, a distinction was made between AI functions and AI content. Usage rights are granted through an open-source license, a proprietary license, or a hybrid model. Particular attention was paid to the issue of contractual prohibitions on certain uses (e.g., reverse engineering, prompt injection). With regard to AI-generated content, however, the problem is that there is generally no copyright protection, meaning that granting usage rights is not possible. In this respect, a solution is provided by contractual provisions, such as restrictions on commercial use, which can reduce liability risks. Obligations to review the output or labeling requirements can also be stipulated in the contract.
With regard to liability, Ms. Mattern discussed product liability law and the new Product Liability Directive. Under Section 1 of the Product Liability Act (ProdHG), liability for defects applies regardless of fault and cannot be excluded by contract. However, fault-based liability for breaches of duty in the use of AI was identified as a problem. In this context, the focus must be on the last causal human action leading to a malfunction of the AI system. In this regard, however, the difficulty of opacity (“black box” problem) persists. Legal uncertainties can be addressed through a contractual solution. In a contract, the distribution of the burden of proof, monitoring and control obligations, logging and reporting obligations, a limitation of liability, or indemnification against third-party claims can be specifically regulated on a case-by-case basis.
The discussion also addressed warranty law with regard to liability for defects. Here, too, supplementary contractual provisions are useful for clarification or to address gaps in the law, such as a specific description of performance or provisions regarding responsibility for the origin and quality of training and test data. Data protection is also regularly a key component of contractual provisions.
Here, for example, clear delineations of responsibilities can be established, or contractual clarifications can be provided regarding the scope of processing, the rights of data subjects, responsibility for disclosure obligations, and so on. It may also be appropriate to include a declaratory statement regarding compliance with data protection requirements or a contractual prohibition on processing certain categories of data.
The protection of trade secrets was also addressed. It was noted that the Trade Secrets Protection Act does not apply automatically, but only when active, appropriate confidentiality measures are in place. The possibilities for contractual safeguards for both the provider and the user were discussed.
The presentation was followed by a lively discussion. Among other things, the discussion focused on the question of how contractual safeguards can be established in light of changes in case law and regulation, and how it can be ensured that AI contract clauses are reviewed and, if necessary, adapted. The speaker also answered numerous questions related to copyright law.
About the Lecture Series "Automated Systems"
Further information about the lecture series "Automated Systems" as well as other retrospectives on past events can be found here.
About the Interdisciplinary Institute for Automated Systems e.V. (RifaS)
The institute was founded in the fall of 2017 and stands for interdisciplinary, national, and international research in various areas of automated systems. The research areas can be divided into traffic and mobility, production and economy, as well as medicine. The institute brings together both scientists and practitioners and is active not only in research but also in teaching.
Information on current events can always be found at www.rifas.de.
