The increasing deployment of artificial intelligence in economic decision-making processes is progressively affecting insolvency-related contexts. The EU AI Act establishes a harmonised regulatory framework across the European Union; while it does not bring about any direct changes to substantive insolvency law, it fundamentally redefines the legal requirements for the use of algorithmic systems in insolvency-related contexts. The Act does not aim to control specific industries but rather follows a risk-based approach that focuses on the potential impact of AI systems on fundamental rights and economic participation.
Insolvency systems are inherently data-driven decision-making processes. The use of AI systems can contribute to increased efficiency and consistency in insolvency-typical areas such as the assessment of creditworthiness, the identification of distress situations, the structuring and prioritisation of claims, the support in business restructurings, and the management of court cases. At the same time, however, there is a risk that algorithmic systems will exert a significant influence on core insolvency-related positions, such as access to financing, the treatment of creditors, or the conduct of proceedings. Thus, when exerting such decisive or materially influential effect, AI applications will frequently qualify as high-risk systems. Against this background, the classification of such AI systems under the EU AI Act is of central importance.
This classification entails comprehensive regulatory requirements. For bankruptcy situations, these include, inter alia, the establishment of a continuous risk management system, ensuring high-quality and non-discriminatory data governance as well as extensive (technical) documentation and logging obligations. These requirements serve to ensure the traceability and legal verifiability of algorithmic processes. Furthermore, the EU AI Act requires sufficient transparency of the systems to enable their output to be understood appropriately. Crucially, effective human oversight must be ensured to prevent the ultimate decision-making responsibility from being shifted to the algorithmic systems. Finally, AI systems must be designed in a way that is secure and robust against attacks or tampering.
In addition, the implementation of these requirements poses specific risks in the context of insolvency systems. On the one hand, compliance with regulatory requirements can entail significant costs which place a particular burden on smaller market participants or companies already in financial distress. This potentially conflicts with the objectives of insolvency law, which is geared towards preserving economically viable structures and ensuring efficient market consolidation. On the other hand, the use of AI systems implicates fundamental principles of procedural fairness, especially in terms of due process and evidentiary contexts, as the deployment of complex, potentially non-transparent systems can jeopardise these requirements if algorithmic assessments effectively become decisive without their decision-making bases being sufficiently verifiable. Added to this is the risk of inconsistent application of the provisions of the EU AI Act and the resulting regulatory fragmentation. Despite its direct applicability, enforcement of the Act is necessary while the corresponding discretion remains with the EU Member States’ authorities. Differing supervisory practices can lead to uncertainty, particularly in cross-border insolvency scenarios, and complicate coordination among the parties involved. A coherent and coordinated application of the Act is therefore of central importance.
Considering these considerations, there is a need for targeted integration of AI regulation and insolvency law requirements. Clear guidance should be provided by the European Commission for the classification of AI systems in insolvency contexts to ensure legal certainty for providers and users, especially in terms of the distinction between purely assistive systems and those exerting material influence on decision-making. Moreover, targeted capacity-building among courts and insolvency practitioners is essential to ensure the informed and critical use of AI systems as well as the implementation of coordination mechanisms across EU Member States. With this, the proper handling of AI systems and the ability to critically evaluate their results can remain effectively safeguarded.
Overall, the EU AI Act fundamentally changes the framework for the use of AI in insolvency-related systems. The key challenge lies in adopting the efficiency gains associated with its use without compromising the core principles of insolvency law, which focus on resolving financial distress efficiently, fairly, and transparently by balancing the interests of debtors, creditors, and the wider economy.