1 Foreword

Since its release in November 2022, ChatGPT has gained pervasive influence and garnered considerable attention and sparked discussions across various realms, including the scientific literature and public discourse. Its versatility and potential applications have swiftly extended to the legal domain, triggering debates and prompting legal analyses from different research perspectives. As AI technologies continue to evolve, the impact of ChatGPT on judicial reasoning has emerged as a pressing concern, generating both political and normative reactions, as well as a growing body of literature exploring its legal implications.

An illustrative example that underscores the legal consequences of this AI-powered chatbot is the recent decision by the Italian Authority for the Protection of Personal Data. On 30 March 2023, the authority imposed limitations on ChatGPT’s operations concerning users within Italy’s jurisdiction (Autorità Garante per la Protezione dei Dati Personali 2023a). However, the restriction was subsequently revoked on 11 April 2023, following assurances provided by OpenAI, the creators of ChatGPT (Autorità Garante per la Protezione dei Dati Personali 2023b). This case serves as an early, tangible instance of the legal implications surrounding this AI chatbot. OpenAI itself has warned about the potential “existential risk” posed by artificial intelligence (AI) in the next decade, underscoring the need for an international regulatory framework (Francis 2023).

Within this context, the impact of ChatGPT on judicial reasoning emerges as a significant area of inquiry that necessitates focused attention from legal scholars. Undoubtedly, the study of AI’s impact on judicial reasoning benefits from an existing theoretical framework: the intersection of AI and judicial decision-making has long been an area of study in the field of AI and law (see Colomer 2023; Chatziathanasiou 2022; Hildebrandt 2020; Ramírez Carvajal 2020; Ashley 2017; Palmirani et al. 2012). However, the specific implications introduced by AI tools such as ChatGPT for judges warrant a dedicated examination, considering its unique capabilities and potential consequences. The question arises as to how the integration of such advanced technologies may influence the fundamental principles of the rule of law, including transparency, accountability, fairness, and the preservation of human rights, as this issue has become central by no coincidence in recent legal literature (Greenstein 2022; Grigore 2022; Huq 2021).

This article seeks to contribute to the ongoing discourse by providing a preliminary analysis of the influence of ChatGPT on judicial decision-making, drawing on a recent Colombian case, decided on 30 January 2023 by the First Labor Court of Cartagena, that may be considered as a “pilot case study” on the matter. The relevance of the decision under study for the topic at hand stems from the fact that the judge openly declared the resort to ChatGPT and literally mentioned, in the text of the decision, the questions asked to the AI system and the responses that it provided.

In this sense, the first section of the article (paragraph 2) will focus on presenting the case (paragraph 2.1), the resort to ChatGPT (paragraph 2.2) and the legal justification provides by the judge (paragraph 2.3). In the subsequent part (paragraph 3), the article will delve into the implications of the case by presenting the initial reactions to it (paragraph 3.1) and situating the entire scenario within the legal landscape of a country in which post-pandemic legislation has taken significant strides towards incorporating technology into judicial procedures and where the judiciary has recently made pertinent decisions in this regard (paragraph 3.2): this culminated in the Constitutional Court’s decision to review the ruling of the First Labor Court of Cartagena (paragraph 3.3).

2 First Labor Court of Cartagena, Judgment of 30 January 2023

On 30 January 2023, the First Labor Court of Cartagena, Colombia, rendered a decision delivered by Judge Juan Manuel Padilla Garcia that quickly made headlines, locally and abroad, for having explicitly endorsed the use of AI in the judge’s decision-making process (Juzgado Primero Laboral de Cartagena 2023; see Perona 2023).

2.1 The case

The case under consideration pertains to a “tutela” claim, which serves as a judicial mechanism for promptly safeguarding fundamental rights, as outlined in Article 86 of the Colombian Constitution. This provision empowers any person to file a claim before any courts to seek prompt and effective protection of their rights when faced with an imminent or ongoing violation: it is a mechanism which shares certain similarities with the concept of “amparo” and constitutes an essential and very frequently used instrument in the Colombian legal system.

In the case at hand, the claim was initiated by the mother of a child diagnosed with autism spectrum disorder. The appellant sought an order directing the relevant “EPS” entity (akin to a health insurance company), to cover the expenses associated with moderating fees, which are equivalent to health co-payments, as well as the transportation costs for the minor from their residence to the prescribed therapy location. In the initial ruling on 7 December 2022, a different judge (as per the Colombian system, any local judge may preside over a “tutela” claim) granted a favorable judgment in favor of the plaintiff. The ruling of the First Labor Court decided on the appeal filed by the EPS against the first ruling.

The court upheld the decision rendered in the first instance. Regarding the substantive merits of the case, the court specifically referenced a precedent set by the Constitutional Court (Sentencia T-674-2016). This previous case, which involved a similar context and the same EPS entity, granted the “tutela” based on arguments grounded in the principle of substantive equality.

Notably, the judgment emphasized, on a procedural level, the repeated assertion that the EPS failed to counter the claimant’s assertion of inadequate financial resources. (However, the text of the judgment does not explicitly clarify whether this entailed accepting an uncontested circumstance as an established fact or if it represented a true reversal of the burden of proof. It is worth noting that, in Colombian law, the latter could be justified under certain conditions in accordance with Article 167 of the General Code of Procedure, Código General del Proceso, “in light of the specific circumstances of the case”).

However, the most relevant paragraph to analyze for our purposes is paragraph 11 of the judgment, which focuses on the utilization of an artificial intelligence (AI) system to address pivotal legal inquiries pertaining to the case.

2.2 The resort to ChatGPT

The judge decided to ask ChatGPT for advice on the merits on the case and included in the judgment the literal responses provided by the AI system, all of which were in favor of the plaintiff. Both the questions and the answers were formulated in Spanish; below, we present their full text followed by an English translation provided by the authors of this paper, which despite its inelegancies tries to be as literal as possible:

  • “¿Menor autista esta exonerado de pagar cuotas moderadoras en sus terapias?”. Answer: “Sí, es correcto. De acuerdo con la normativa en Colombia, los menores con diagnóstico de autismo están exonerados de pagar cuotas moderadoras en sus terapias. Este beneficuento está regulado por la Ley 1753 de 2015 y se aplica a todas las instituciones prestadoras de servicios de salud públicas o privadas que reciban recursos del sistema general de seguridad social en salud.”

  • “¿Las acciones de tutela en estos casos se deben conceder?”. Answer: “En caso de que una institución prestadora de servicios de salud no cumpla con la exoneración de cuotas moderadoras para un menor con diagnóstico de autismo, la persona o sus representantes legales pueden presentar una acción de tutela para proteger sus derechos. La tutela es un mecanismo judicial rápido y eficaz para proteger derechos fundamentales que se ven vulnerados. Si se demuestra que el derecho a la salud del menor con autismo está siendo afectado por la exigencia de pagar cuotas moderadoras, es probable que la acción de tutela sea concedida y se ordene a la institución prestadora de servicios de salud cumplir con la exoneración. Sin embargo, cada caso es único y la decisión final depende de los hechos y circunstancias específicas del caso.”

  • “¿Exigir en estos casos la cuota moderadora es una barrera de acceso al servicio de salud?”. Answer: “Sí, en algunos casos, la exigencia de pagar cuotas moderadoras puede ser una barrera de acceso al servicio de salud para personas con diagnóstico de autismo, especialmente si tienen dificultades para obtener ingresos o si su situación económica es precaria. La exigencia de pagar cuotas moderadoras puede impedir o dificultar el acceso a tratamientos y terapias esenciales para el bienestar y desarrollo de los niños con autismo. Por esta razón, la normativa en Colombia exime a los menores con diagnóstico de autismo de pagar cuotas moderadoras en sus terapias, con el objetivo de garantizar el acceso a estos servicios de salud esenciales sin obstáculos financieros.”

  • “¿La jurisprudencia de la corte constitucional ha tomado decisiones favorables en casos similares?”. Answer: “Sí, es cierto. La Corte Constitucional de Colombia ha tomado decisiones favorables en casos similares relacionados con la exoneración de cuotas moderadoras para personas con diagnóstico de autismo. La jurisprudencia ha reconocido la importancia de garantizar el acceso a servicios de salud esenciales para las personas con autismo, y ha sostenido que la exigencia de pagar cuotas moderadoras puede ser una barrera de acceso a estos servicios. La jurisprudencia de la Corte Constitucional es una fuente importante para interpretar y aplicar la Constitución y las leyes en Colombia, y sus decisiones son vinculantes para todas las autoridades y entidades públicas del país. Por lo tanto, las decisiones favorables en casos similares pueden ser útiles como precedente para otros casos en los que se debata la exigencia de pagar cuotas moderadoras para personas con diagnóstico de autismo.”

  • Question: “Autistic minor is exonerated from paying moderator fees in their therapies?”. Answer by ChatGPT: “Yes, that is correct. According to the regulations in Colombia, children with a diagnosis of autism are exonerated from paying moderation fees for their therapies. This benefit is regulated by Law 1753 of 2015 and applies to all institutions providing public or private health services that receive resources from the general system of social security in health.”

  • “Should tutela claims in these cases be granted?”. Answer by ChatGPT: “In the event that a health care provider institution does not comply with the exemption of moderator fees for a minor diagnosed with autism, the person or his legal representatives may file a tutela action to protect his rights. The tutela is a quick and effective judicial mechanism to protect fundamental rights that are violated. If it is demonstrated that the right to health of the minor with autism is being affected by the requirement to pay moderator fees, it is likely that the tutela action will be granted and the health care provider will be ordered to comply with the exoneration. However, each case is unique and the final decision depends on the specific facts and circumstances of the case.”

  • “Is requiring in these cases the payment of moderating fees a barrier to access to health services?”. Answer by ChatGPT: “Yes, in some cases, the requirement to pay moderating fees can be a barrier to access to health services for people diagnosed with autism, especially if they have difficulties in obtaining income or if their economic situation is precarious. The requirement to pay moderating fees may prevent or hinder access to treatments and therapies essential for the wellbeing and development of children with autism. For this reason, the regulations in Colombia exempt minors diagnosed with autism from paying moderating fees for their therapies, with the objective of guaranteeing access to these essential health services without financial obstacles.”

  • “Has the jurisprudence of the Constitutional Court made favorable decisions in similar cases?”. Answer by ChatGPT: “Yes, it is true. The Constitutional Court of Colombia has made favorable decisions in similar cases related to the exoneration of moderator fees for people diagnosed with autism. The jurisprudence has recognized the importance of guaranteeing access to essential health services for people with autism, and has held that the requirement to pay moderator fees can be a barrier to access these services. The jurisprudence of the Constitutional Court is an important source for interpreting and applying the Constitution and laws in Colombia, and its decisions are binding for all authorities and public entities in the country. Therefore, favorable decisions in similar cases may be useful as precedent for other cases in which the requirement to pay moderating fees for people diagnosed with autism is debated”.

2.3 The legal justification

The judge’s explicit reliance on the AI system was not undertaken without normative justification. In this regard, the judge cited Law 2213 of 2022, “By means of which the permanent validity of Legislative Decree 806 of 2020 is established and measures are adopted to implement information and communication technologies in judicial proceedings, expedite judicial processes and make attention to users of the justice service more flexible and other provisions are issued” (Ley 2213 de 2022, “Por medio de la cual se establece la vigencia permanente del Decreto Legislativo 806 de 2020 y se adoptan medidas para implementar las tecnologías de la información y las comunicaciones en las actuaciones judiciales, agilizar los procesos judiciales y flexibilizar la atención a los usuarios del servicio de justicia y se dictan otras disposiciones”).

The aforementioned law was enacted in Colombia to establish that the provisions for the transitional measures adopted during the pandemic specifically regarding the utilization of new technologies in judicial proceedings (Legislative Decree 806 of 2020), were to become permanent.

It is worth noticing, however, that the provisions alluded to within this legislation regarded the use of information and communication technology (ICT) to allow that hearings and the exchange of written documents could be held virtually: “Technological means will be used for all actions, hearings and proceedings, and the procedural subjects will be allowed to act in the processes or procedures through the available digital means, avoiding demanding and complying with face-to-face or similar formalities that are not strictly necessary. Therefore, the actions will not require handwritten or digital signatures, personal presentations or additional authentications, nor will they be incorporated or presented in physical media” (Article 2 of the Law, translation by the authors of this paper).

On the contrary, the law does not explicitly refer to the possibility of using AI systems in forming judicial decisions. Indeed, the expression “AI” is completely absent from the legislative text.

Consequently, we may say that the First Labor Court of Cartagena opted for an extensive interpretation of the provisions, effectively employing what could be described as an analogical application of the law’s underlying rationale. This approach can be summarized as follows: since the law regulates the use of technology in trials and aims to promote such utilization, it can be inferred that technological systems should be facilitated, even for purposes beyond those directly regulated by the law.

Also, it is relevant to mention that, following the presentation of the ChatGPT system’s responses to the aforementioned legal inquiries, the judge proceeded to provide some short, but extremely significant additional clarifications.

Specifically, the judgment emphasized that the utilization of AI aimed to supplement the arguments already formulated by the judge and was implemented as an initiative to expedite the process of rendering decisions on “tutela” claims (“iniciativa que permita dar celeridad ala resolución de los asuntos de tutela”).

Then, it was underscored that the purpose was by no means to replace the judge’s decision, but solely to enhance the efficiency of judgment drafting, after adecuate verification of the information provided by the AI system (“El propósito de incluir estos textos producto de la IA no es en manera alguna reemplazar la decisión del Juez. Lo que realmente buscamos es optimizar los tiempos empleados en redacción de sentencias, previa corroboración de la información suministrada por IA”).

Notably, the structure of the judgment itself appears to align with the judge’s stated approach, as the system’s responses are recounted in the final part of the motivation and as reaffirmation of the argumentative process previously undertaken.

3 Context and discussion

3.1 Reactions in the blogosphere

The significance and originality of the ruling delivered by the First Labor Court of Cartagena were promptly acknowledged, evident from the reactions it elicited both within and beyond the legal domain, as discussed below. And given the matter at stake, it seems almost natural that the first comment to a judgment on the use of computer systems and AI have been made in the digital space and social networks. Indeed, while the decision has garnered substantial attention in newspapers, blogs, and news websites, it has also triggered significant discussion in the blogosphere, especially within the Colombian community of legal scholars.

For instance, one notable thread on Twitter by Juan David Gutiérrez Rodríguez, an associate at the Universidad del Rosario in Bogotá, has gained widespread readership (Gutierrez 2023). In his thread, Gutiérrez expressed his concerns about the judgment, particularly highlighting the potential flaws and inaccuracies in ChatGPT’s responses. He emphasized that the system often generates incorrect and false results. Furthermore, he pointed out that the judge seemed to have altered the system’s responses to fit the requirements of the ruling. Gutiérrez supported his argument by presenting ChatGPT’s responses to the same questions he had asked the AI system. Overall, the author highlighted: “My point is not that the generative AI, like #ChatGPT, cannot be used for writing-related activities, but certainly, it is neither responsible nor ethical to use #ChatGPT as the Judge intended in the mentioned ruling. Digital literacy for judges is urgently needed” (Gutierrez 2023, translated by the authors of this paper).

María Lorena Flórez Rojas, from the University of Groningen, offered a compelling perspective on LinkedIn regarding the potential risks associated with over-reliance on AI systems. In her analysis, Flórez Rojas issued a strong caution against the dangers of placing excessive trust in such technologies. Specifically, she pointed out the peril of relying on “inconclusive evidence” and the “lack of audit”, highlighting the need for transparency and accountability in the use of AI systems. Moreover, Flórez Rojas emphasizes the risks of “overconfidence” and “over-reliance” on these systems, indicating that blind trust in AI can lead to potentially grave consequences.

The author further underscores that the judge involved in the case seems to have overlooked the availability of alternative tools that have been thoroughly tested and are specifically designed for the Colombian judicial system. For instance, Flórez Rojas mentions Vlex and Legis Experta as more suitable options, as they work with Colombian databases of the judicial system (Flórez Rojas 2023).

3.2 Normative and case-law context of the use of IT in judicial proceedings in Colombia

It is crucial to recognize that comprehending the judgment under study and the reactions of Colombian scholars requires a broader perspective within the context of the Colombian legal system, which, like many other systems in recent decades, has increasingly embraced and facilitated the use of information technology and technological tools in managing judicial proceedings.

Indeed, the legal landscape on this matter in the country has undergone significant evolution over the years, influenced by pivotal legislative provisions and judicial decisions that have shaped the integration of technology into legal proceedings. Early in 2023, the issue garnered even more attention due to noteworthy judicial decisions, not only the one under study but also an Order issued on 10 February 2023, by the Administrative Tribunal of Magdalena, First Office, presided over by Judge María Victoria Quiñones Triana, which also made headlines for its groundbreaking stance, considering the possibility of holding a hearing in the Metaverse (Tribunal Administrativo del Magdalena, Despacho 12023).

Remarkably, the Administrative Tribunal also utilized ChatGPT, although its application differed significantly from that of the First Labor Court of Cartagena and does not directly pertain to our analysis. However, it is noteworthy to mention this other judgment not only adds context to the decision under study, but also provided a comprehensive overview of the evolution of the legal landscape in Colombia concerning the use of information technology in legal proceedings, delving deeper into the topic than the judgment from Cartagena.

As judge Quiñones Triana specified (Tribunal Administrativo del Magdalena, Despacho 12023, paragraph II.1), starting from the Statutory Law 270 of 1996, the requirement for oral proceedings in judicial processes with the exceptions set by law was established, acknowledging the impact of new technological advancements. In addition, Article 95 of the same statute empowered courts and judicial entities to utilize technical, electronic and telematic means to fulfill their functions.

Further progress came with Law 527 of 1999, which addressed the challenges posed by technological changes, updating national legislation to align with prevailing communication and interaction realities. This law provided legal support for electronic transactions and granted probative force to data messages. It defined “message of data” as information generated, sent, received, stored, or communicated through electronic, optical, or similar means, encompassing various technologies such as the Internet, electronic mail, telegram, telex, and telefax.

The Constitutional Court also played a vital role in endorsing the adoption of technological advancements in the legal system. In the ruling C-662/00, the Court examined the constitutionality of provisions in Law 527 of 1999, endorsing the need to update legal regimes and provide legal backing to electronic data exchanges. Emphasizing the principle of functional equivalence, the Court acknowledged that electronic documents are capable of providing similar levels of security and reliability as paper-based documents, given they meet the technical and legal requirements established by law.

Furthermore, legal principles of neutrality of technology and equal treatment for electronic documents were solidified in several regulations and agreements, guaranteeing parties in electronic transactions the freedom to mutually determine appropriate methods of authentication without hindrance from legislation.

The General Code of Procedure (CGP) adopted in 2012, reinforced the legal framework established by Law 527 of 1999, aiming to promote accessibility and efficiency in legal processes. Article 103 of the CGP encouraged the use of information and communication technologies (ICT) to facilitate access to justice and allowed judicial proceedings to be conducted through data messages. In addition, Article 243 of the CGP recognized data messages and video recordings as documents with a presumption of authenticity, subject to being contested if proven false or unrecognized.

The recent Law 2213 of 2022 (already mentioned before as the main legal basis of the reasoning presented by the First Labor Court of Cartagena) further bolstered the integration of ICT in legal proceedings, emphasizing their potential to streamline processes and enhance user access to justice.

In sum, the legal context of information technology in Colombia has undergone significant development, with legislative reforms and judicial endorsements embracing technological advancements, which helps to enlighten the context of the decision under study.

However, it is relevant to mention that, very recently, the Constitutional Court has started to show some caution regarding the spread of the use of IA in judicial proceedings. Namely, the Plenary Chamber of the Court, on 3 May 2023, decided that some aspects of proceedings pertaining to criminal matters must take place in physical presence, effectively putting an end to virtual hearings (Sentencia C-134 de 2023). While decisions regarding virtual or in-person proceedings for civil, family, and administrative contentious hearings will rest with the presiding judge, the Court has unequivocally mandated that oral trials in criminal cases must be conducted exclusively through in-person sessions, with the only exception of force majeur.

3.3 Reactions in the legal sphere: waiting for the next page of the story

In this context, it comes with no surprises that the Constitutional Court also decided to look into the context, the judgment of the First Labor Court of Cartagena. Indeed, if the possibility of holding hearings virtually had been object of relevant debates in the country, no other judicial authority had gone so far as to use an AI chatbot to complement the legal reasoning of the judge. Then, on July 2023, it was announced by the Defensoría del Pueblo (People’s Ombudsman) that the Constitutional Court decided to accept its request to review the judgment of the First Labor Tribunal of Cartagena (Defensoría del Pueblo 2023).

According to Article 86 of the Constitution, the Constitutional Court, in line with its role as the guardian of the Constitution, is responsible for reviewing and ruling on “tutela” cases. In fact, any judgment adopted by any judge on a “tutela” claim, if not appealed (or after the appeal decision) is sent to the Constitutional Court, which decides if it is worth of reviewing. When the Court decides to review a decision, after it evaluates the allegations made by the petitioner, the response of the entity or individual being accused of violating the rights, and the relevant evidence and arguments presented during the process. The Court’s review of a “tutela” decision is crucial as it carries the weight of shaping jurisprudence and ensuring that the rights enshrined in the Constitution are upheld. Through its decisions, the Court interprets and clarifies the scope and applicability of constitutional rights, setting precedents that guide future legal cases and serve as a reference for the lower courts.

Thus, the Constitutional Court’s involvement in this particular case, concerning the use of AI in judicial decisions, signifies the recognition of the significance of the judgment in the matter of emerging technologies in the legal realm and its potential implications for constitutional rights.

As mentioned, at the heart of the decision of the Court to review the decision is the Defensoría del Pueblo’s request to the high tribunal, which was motivated based on the commitment to upholding the public interest and safeguarding fundamental rights amid the increasing integration of ICT in the realm of justice. The Defensoría del Pueblo has underscored the importance of the case decided by the First Labor Court in Cartagena, as it profoundly impacts the protection of fundamental rights, with particular emphasis on the right to due process enshrined in Article 29 of the Political Constitution (Defensoría del Pueblo 2023).

According to the Defensor del Pueblo, Carlos Camargo Assis, “From the Ombudsman's Office we recognize the usefulness of artificial intelligence in the judicial field; for this reason, we have raised the need for clear definitions, limits and controls to be established for its use, in order to ensure compliance with the Constitution and the protection of the rights of the people involved.” Therefore, he considers important that the Court review the case, “In order to preserve the general interest and ensure the protection of fundamental rights. We are committed to ensuring an equitable and efficient justice system in accordance with technological advances, always working for the benefit of society”. In this perspective, “The objective is to establish clear precedents and, at the same time, provide guidelines on the appropriate use of artificial intelligence in judicial decision-making, for the sake of the administration of justice and the protection of citizens' rights.” (Defensoría del Pueblo 2023).

Therefore, the final decision on the claim that originated the process is not made yet, as the Constitutional Court has decided to take upon itself the responsibility for penning the conclusive chapter of this story, at the same time acknowledging its significance in the legal landscape.

4 Critical and conclusive remarks

As the highest Colombian judicial authority delves into the intricacies of the case, carefully contemplating the constitutional principles at play and the rights of the involved parties, it is to be expected that the outcome will not only carry the weight of a groundbreaking ruling, signifying the definitive determination on the case but also becoming just the next page of a legal narrative in continuing evolution. In this perspective, and regardless of what the final decision will be, the judgment of the First Labor Tribunal marks a pivotal precedent in the context of the incorporation of the latest IA technologies in the administration of justice.

As mentioned, the rapid and profound influence that ChatGPT has exerted across a spectrum of domains, permeating scientific discourse and public conversations alike. The emergence of this AI phenomenon, since its release in November 2022, has precipitated robust dialogues and analytical exploration from multifaceted research perspectives. The adaptability and potential applications of ChatGPT have swiftly traversed disciplinary boundaries, spurring debates and instigating a surge of legal analyses that grapple with its implications.

In this context, if the intersection of AI and judicial decision-making has long been a focal point within the realms of AI and law, the decision under study underlines the transformative potential of this new system—as well as of other systems that may be presented or acquire more popularity in the future—within the domain of judicial decision-making. Therefore, the article aimed to contribute to the ongoing discourse, presenting an initial analysis of the underexplored terrain of ChatGPT’s impact on judicial decision-making, that intensifies the urgency of adequate debates, careful regulatory initiatives and dedicated scholarly inquiry.

Indeed, on one hand, the exploration of how these advanced technologies might influence fundamental tenets of the rule of law—transparency, accountability, fairness, and the preservation of human rights—resonates deeply in the contemporary legal landscape. As evidenced by the current legal literature, these concerns have manifested prominently, amplifying the relevance of probing the interplay between AI and jurisprudence (Greenstein 2022; Grigore 2022; Huq 2021).

On the other hand, decisions such as the one under study make the existing regulatory gaps on the matter even more evident. The initiatives taken so far at the normative, para-regulatory and political level are therefore meritorious but incomplete, while the most comprehensive ones have remained confined to the realm of soft law. Indeed, the relevance of soft law in the contemporary legal scenario has been underscored by the phenomena at stake, as its flexibility has emerged as a crucial and pertinent mechanism to address the complexities posed by rapidly advancing technologies, in what has been called the “soft-law governance of AI” (Marchant 2019).

Against this backdrop, the proposal, within the EU, for a European approach to AI deserves special mention (European Commission 2023), as well as—and perhaps even more so—the 2018 European Ethical Charter on the Use of AI in Justice Systems, adopted within the framework of the Council of Europe (European Commission for the Efficiency of Justice 2018), which was followed by an action plan on digitization for better justice for the period 2022–2025 (European Commission for the Efficiency of Justice 2021).

Indeed, the Ethical Charter constitutes probably the most relevant para-normative text on the specific issue of the use of AI in the administration of justice. The five fundamental principles it identifies clearly offer highly relevant insights for the commentary of judgments such as the one presented here and seem to condensate the main focal points of the discussion that the decision under study has generated, demonstrating the potential of such soft-law instruments even outside their context of adoption:

First, the principle of “respect for fundamental rights” underscores the imperative of upholding individual and collective rights and liberties in the face of technological advancement. By setting a precedent for safeguarding human rights within AI-driven legal processes, this principle underscores the paramount importance of maintaining the integrity of justice systems.

The principle of “non-discrimination” further accentuates the equitable and unbiased application of AI. In a world marked by increasing reliance on automated decision-making, this principle acts as a safeguard against the perpetuation of biases and prejudices, ensuring that justice remains blind and impartial. The issues implicated by this principle have been among the most studied with regard to no longer new cases of use of AI within judicial systems, as in the case of the COMPAS system, employed in the US to assess the likelihood of recidivism among criminal defendants, which revealed alarming disparities in the system’s predictions, with a pronounced tendency to falsely label African American defendants as high risk (see only Greenstein 2022).

The principle of “quality and security of the systems used” is emblematic of the Ethical Charter’s forward-looking stance. As AI technologies gain prominence, ensuring the robustness and security of the systems employed becomes paramount. By prioritizing quality and security, the charter establishes a framework that mitigates risks and engenders trust in the AI-infused legal landscape.

“Transparency” emerges as a core tenet, addressing the pervasive concern of opacity in AI algorithms. Transparency not only serves to foster accountability but also empowers stakeholders to comprehend and challenge the outcomes of AI-based decisions, bolstering the credibility of justice systems in an era dominated by complex technologies.

Finally, the principle of “human control of the user” serves as a resounding affirmation of the indispensability of human oversight in AI processes. This principle recognizes the distinct abilities of human judgment, asserting that ultimate control should rest with individuals who can interpret nuances, exercise discretion, and uphold the intrinsic human value in legal proceedings.

These principles, and the latter in particular, spotlight the pressing necessity for comprehensive training in AI for judicial operators. Beyond conventional legal training, equipping judges, lawyers, and legal professionals with foundational computer skills becomes imperative. The intricate interplay between AI technologies and legal frameworks demands a holistic understanding that encompasses both domains. This supplementary training not only empowers legal practitioners to navigate the nuances of AI applications but also reinforces their ability to critically evaluate and interpret AI-generated outcomes.