My AI, my code, my secret – Trade secrecy, informational transparency and meaningful litigant participation under the European Union's AI Liability Directive Proposal

Abstract

In European Union (EU) law, the AI Liability Directive (AILD) proposal included a right for victims of harm caused by high-risk AI systems to request the disclosure of relevant evidence. That right is, however, limited by the protection of trade secrets. During legal proceedings, business confidentiality can indeed restrict the victims’ access to evidence, potentially precluding them from fully understanding the disputed facts and effectively making their views known before a court. This article examines whether the AILD provided sufficient procedural mechanisms to ensure that litigants can effectively participate in judicial proceedings, even when critical evidence is withheld from them, due to legitimate trade secret protections. Our analysis draws on the evidentiary challenges highlighted in emerging global AI liability cases and selected CJEU case law, which provide guidance on how a balance can be struck between legitimate confidentiality and a workable level of informational transparency, necessary for an informed and fair resolution of future AI liability disputes.