ARTIFICIAL INTELIGENCE AS MEANS TO A JUDICIAL FUNCTION ALIGNED WITH ACCESS TO JUSTICE
DOI:
https://doi.org/10.17058/rdunisc.vi78.21009Abstract
The embedment of artificial intelligence systems into judicial functions, although often presented as an efficiency-enhancing measure, promotes tension with the constitutional guarantee related access to justice. This paper examines, through theoretical and documentary research, whether such applications aiming to grant celerity in ruling might deviate jurisdiction from access to justice as a material parameter of judicial legitimacy. The research has as object the intersection between the constitutional guarantee named access to justice and the inherent limitations of AI mechanisms. The objective is to indicate minimal components of a governance structure applicable to AI systems used as support to the exercise of judicial function. The analysis demonstrates that algorithmic mechanisms, grounded in statistical inference and trained on data sets shaped by historical selectivity, tend to replicate rationalities that the justice system ought to overcome, not automate. It concludes that the opacity of these models, combined with their epistemic limitations for tasks requiring contextual deliberation and value-based reasoning, recommends robust governance, with meaningful human oversight, and effective adherence to the constitutional foundations that structure the provision of jurisdiction.
