Recoding Criminal Justice in the Age of Generative AI Disruption

Descripción del proyecto

Cómo la inteligencia artificial generativa está transformando la justicia penal

Es posible que los agentes de policía estén utilizando la inteligencia artificial (IA) para redactar informes. Es posible que los jueces la utilicen como referencia a la hora de dictar sentencia. Es posible que se evalúe a los reclusos mediante algoritmos que predicen su comportamiento futuro. La IA generativa (IAG) está transformando silenciosamente la justicia penal, a veces a través de canales oficiales y otras a través de profesionales que utilizan estas herramientas por iniciativa propia. El equipo del proyecto RecodingJustice, financiado por el Consejo Europeo de Investigación, analiza esta transformación a lo largo de todo el proceso penal, desde la actuación policial hasta la investigación, el juicio, el encarcelamiento y la rehabilitación. En el proyecto, que se analizan cuatro países (Brasil, Israel, los Países Bajos y los Estados Unidos), se examinan tanto lo que las instituciones adoptan formalmente como lo que los profesionales utilizan de manera informal entre bastidores. El objetivo: un nuevo marco y estrategias normativas para proteger los derechos humanos y garantizar la equidad de los procedimientos, a medida que la IA va transformando la forma en que se imparte justicia.

Objetivo

How will Generative AI (GenAI) reshape criminal justice? Unlike earlier AI systems, often confined to specific tasks and users, GenAI democratizes and boosts AI’s capabilities, transforming our understanding of AI’s role in justice systems. GenAI enables wider access—for police officers, investigators, lawyers, judges, prison staff, and rehabilitation workers—and diverse applications, from aiding in crime analysis and drafting legal arguments to generating personalized rehabilitation plans, spanning all stages of the justice cycle, from pre-crime prevention to post-incarceration reintegration. Its potential to enhance efficiency, personalize interventions, and expand access to justice is vast. Perhaps most strikingly, GenAI excels in tasks typically human-dominated, such as interpreting complex narratives or creating nuanced analyses. These advancements offer remarkable opportunities for justice but also carry substantial risks that extend beyond the familiar risks of bias, opacity, and lack of accountability. While those dangers continue to shape the AI landscape, GenAI gives rise to novel perils, such as hallucinations and large-scale manipulation, like systemic exploitation or misuse. GenAI’s risks are amplified by its dual use: institutionalized applications governed by inadequate regulation and informal uses that often largely operate “under the radar,” allowing injustices to proliferate, affecting both individual outcomes and the broader integrity of the criminal justice system.

Recoding Justice will systematically examine both formal-institutionalized and informal-ad hoc uses of GenAI and their impact internationally across four diverse and distinct legal settings (Brazil, the United States, the Netherlands, and Israel). The project will draw on empirical case studies to reimagine justice in the AI age on the theoretical level. The proposed theory could guide regulatory strategies to ensure that GenAI’s disruptive potential aligns with justice principles.