The Ethical Use of AI Tools in the Judiciary
Geraldo Dutra de Andrade Neto, Executive Secretary Judge of the International Union of Portuguese Language Judges (UIJLP) and Secretary of International Relations of the Brazilian Magistrates Association (AMB)
Increasingly present in people's daily lives, Artificial Intelligence (AI) tools have brought about essential changes – from the job market to the fields of health and education, to name a few examples – with great potential to also impact the Judiciary. There is a promise of efficiency and speed embedded in new technologies, which, however, do not come without risks. The question is: how to make the most of these tools for the daily work of judges while ensuring respect for fundamental rights and guarantees? This is not just a technical dilemma, but an ethical debate that requires deep reflection. After all, in the case of Justice, the pursuit of productivity cannot eliminate the analysis of nuances and contexts only perceptible to the judge's eye.
The history of Law is marked by continuous evolution: if, in the beginning, the challenge was to apply the legislation inscribed on clay tablets with precise chisels, the problem now is to interpret laws and deliver justice in a scenario where new information technologies, gradually and inexorably, become ubiquitous, performing functions previously reserved for humans. As advanced as they may be, AI tools do not ponder – and much less understand in a human way, with feelings – the practical consequences of a judicial determination in people's lives. They may even suggest text drafts, after countless calculations and information cross-references, but they lack the essential: the discernment that only a life of tangible experiences can provide.
With this concern in mind, the International Union of Portuguese Language Judges (UIJLP), gathered in Ordinary General Assembly on the first day of November, approved the "Foz do Iguaçu Charter" (in reference to the city in Paraná, Brazil, that hosted the event), which outlines guidelines for the use of AI tools in Justice. More than an action protocol, the manifestation of ethical principles inaugurates a line of prevention against the virtual dehumanization of the Judiciary, establishing that AI technologies, although they constitute auxiliary apparatus, will never replace judges. In other words: they should never enjoy autonomy, even if feasible, to produce decisions in ongoing cases.
AI tools should contribute to the independence of the Judiciary, operating as a support device, not as a second agent or parallel judge. Every text suggested by AI needs to be reviewed, altered, analyzed, and weighed by the judge to verify its compatibility with the judge's ruling. In a system that values impartiality in the distribution of justice, the verdict – which has the power to profoundly transform the reality of citizens – must remain in human hands, reaffirming the non-delegable nature of jurisdiction.
Our communication has personalized AI technologies – and the way we express ourselves is relevant. AI technologies in Justice are work tools that bring benefits and risks, but we should not treat them as sentient. The risk is losing the human dimension of judicial activity, which is heightened by the use of new technologies. Although they may suggest acceptable texts for concrete cases, they do not replace the need for a judge who judges with empathy and humanism.
Trust in the integrity of judges' work must always be preserved. The complexity of AI algorithms, understandable to few, makes it controversial to require judges to inform in their decisions about the use of AI tools. Besides being difficult to comply with and verify in daily practice, it will certainly give rise to new appeals, which will inevitably delay the conclusion of cases. Such a measure can undermine public confidence in the justice system as a whole, on which the independence and preservation of judges' legitimate authority depend. The most important thing is to ensure, through ethical principles and legal norms – already existing in the Brazilian case – that the judicial function will not be delegated to any automated systems.
The Foz do Iguaçu Charter does not ignore, however, the gains from AI tools. The proper use of new technology offers promptness in examining immense volumes of data, identifying patterns, and rationalizing procedures, with a view to reducing errors – benefits that increase the quality of judicial service. It is important to emphasize, as stated in the Charter, the need to preserve the Judiciary's autonomy in developing its own solutions, since Justice cannot become dependent on artifacts controlled by other powers, under penalty of weakening itself in the face of undue pressures and undermining judicial independence.
The position of the UIJLP ultimately demonstrates that, although AI technologies are valuable to assist Justice in fulfilling its functions, they do not replace human evaluation – absolutely essential to maintain public trust, legal security, and the credibility of the Judiciary. The key, therefore, is ethics. The judicial decision, with all its complexity, is only rendered by the judge, who has the capacity to weigh factors that go beyond the cold and impersonal reach of lines of code.
