Emma Lunardi & Zoé Schwindt, Étude de l’admissibilité de la preuve immersive partiellement générée par l’IA au Québec et à l’international, 2025, 18p., Document de travail n°39
In this report, we explore the admissibility of immersive 3D evidence, partially generated by artificial intelligence, in legal proceedings in Quebec and abroad. This is part of our project to implement immersive 3D representations of evidence in virtual reality headsets, in order to improve the conduct of trials. Although the concept of immersive evidence has been around for some forty years, in the past it took more the form of visual representations designed by parties and experts to illustrate their facts. Today, with the 3D generation of evidence and its possible alteration by AI, we want to go one step further.
The study begins by examining the system for admitting evidence in Quebec, particularly in light of the principles governing electronic evidence contained in the Loi concernant le cadre judiciaire des technologies de l’information (LCCJTI). Various criteria are presented, notably for assessing the authenticity, integrity and reliability of documents presented as evidence. Faced with these different criteria, we then discussed how these rules could be satisfied by immersive evidence partially generated by AI. A number of challenges were raised, including the transparency of the AI used to generate potentially missing parts of the evidence in the resources, as well as issues of bias and manipulation of the evidence.
Then, we looked at foreign regimes for admitting evidence, in particular those of the USA and the UK, to determine whether they are more inclined to admit immersive evidence partially generated by AI. In the latter, we distinguished different levels of evidence being subject to different admissibility criteria, a system that might more readily admit our evidence. However, we noted that, despite features such as these being of great interest in our study, issues remain concerning the reliability of AI in general, the probative force that such evidence can have, and the lack of studies carried out on the subject.
DOCUMENT
This content has been updated on 04/11/2025 at 11 h 41 min.
