Artificial intelligence and marriage nullity proceedings: a dangerous combination?

Abstract

This essay addresses the deployment of artificial intelligence in canonical marriage nullity proceedings, examining the doctrinal dialectic between the restrictive approach, which confines AI to merely ancillary functions, and the expansive one, aimed at leveraging its generative capabilities as cognitive support for the judge. Both positions converge on the principle that AI may operate solely as an Auxilii Instrumentum, without supplanting the moral certitude and free evaluation reserved to the human judge. The study proceeds with an experimental phase conducted on an anonymised real case of exclusion of indissolubility, submitted to two AI systems: Canonicus AI, a GPT specialised in canonical matrimonial law, and NotebookLM, a general-purpose RAG system. Five tasks were assigned to both: analysis of evidentiary coherence, identification of instructional gaps, assessment of advocacy briefs and Animadversiones, formulation of the votum, and counter-argumentation. Results demonstrate pertinent, hallucination-free, non-counterepistemic outputs consistent with Rotal jurisprudence. The author concludes that the AI-canonical process pairing poses no danger and argues that ecclesiastical tribunals may serve as a privileged laboratory for governing AI risks within an authentically Christian anthropological framework.