Estado de Derecho
Digital Justice: A Road of No Return
Something is deeply changing the Mexican justice system. For too long it went unnoticed, but it is starting to show.
Last month, during the XVIII Ordinary General Assembly of the Mexican Association of Justice Providers (AMIJ), which revolved around the theme "The challenges for a digital justice in Mexico", the president of the Federal Judiciary, Minister Norma Piña, highlighted the need for judges to look forward to the future. An incentive to reflect on how new technologies can expand access to justice, improve case management, make the work of justice system operators more efficient, reduce costs and explore new forms of online dispute resolution. As well, she also underscored the importance of integrating these technologies into the work of the judiciary in a responsible and ethical fashion.
The Transparency in Justice Program from México Evalúa finds it positive news that the topic is given relevance; since we have been analyzing for several years the enormous potential of new information technologies to increase access to justice, reduce resolution times, the backlog of courts and tribunals, and prevent corruption. However, as we pointed out in our Guidelines for Digital Transformation in the Administration of Justice, these processes can only be successful if they have the strong will of the highest authority of the body that intends to promote them, since internal resistance and the obstacles to overcome along the way are usually significant.
Minister Piña's words and the fact that the AMIJ has dedicated its annual assembly to this topic (which resulted from the vision of its executive secretary, Judge Armando Maitret) are not the only positive signs. There is also the Mexican Bar Association, which dedicated its own National Congress in 2022 to the same topic. So far, it seems that the main actors needed for such transformation to be successful are aligned.
The next turn of the screw: artificial intelligence
While digital justice is already a reality in much of the country, (in our Diagnosis of the implementation of technological tools in the judiciary in Mexico, we were able to detect great progress at the local level) the integration of artificial intelligence (AI) in the administration of justice still remained a pending issue until quite recently. In the framework of the AMIJ Assembly, I had the honor of participating, along with leading experts in the field, in a roundtable discussion dedicated to how AI impacts the way justice is delivered.
Certainly, there are different types of AI applications in justice, most of which seek to streamline and make the work of justice operators -judicial officials, prosecutors, defenders and lawyers- more agile and efficient. The following can be enlisted:
- Chatbots or applications to guide people and give them basic legal advice. They have been developed by legaltech companies and also courts that seek to guide citizens and allow them to access their services without the need for a legal representative or offer alternatives to the judicial process. For example, the case of the Civil Resolution Tribunal of British Colombia in Canada. The Tribunal has developed a solution explorer that allows people to resolve conflicts related to small claims or disputes between condominium owners or associations. The explorer asks users questions to propose solutions to their problem, seeking as a priority for the parties to negotiate or, alternatively, offering them a conciliator to help them reach an agreement. Only if they fail to do so, the matter is brought to court and is completely resolved remotely.
- On the other hand, systems have been developed, by both law firms and judicial institutions - such as the Constitutional Court of Colombia, for example, with PretorIA – to make possible to analyze and classify large amounts of text, summarize documents and/or identify patterns, saving hours of work. The precondition for such systems to work is that the documents to be analyzed are presented in electronic format.
- Support systems have also emerged for the identification of useful case law, which can be used by litigants, judges and/or academics, such as the new JusticIA chatbot, that supports users in their queries to the Weekly Judicial of the Federation.
- There are also predictive AI tools to support the decisions of certain actors in the justice system. Some of them aim to support judges in making decisions on the precautionary measures to be taken when subjecting the defendant to a proceeding, by predicting his risk of absconding, or when the sentence is to be applied, by assessing his risk of recidivism. These systems, however, have been questioned, as they tend to amplify the socio-racial biases present in the antecedents of former databases used to train the algorithms.
- Other tools have been developed by large firms or businesses that offer their services to these firms to calculate the likelihood of winning a lawsuit, particularly when it is known who the judge will be, or the likely amount that can be possibly obtained by negotiating. While these tools can be useful in guiding clients in their decision as to whether or not to pursue a lawsuit or seek a potential settlement, they still present their own complications. Firstly, some people may find it difficult to find a lawyer who will agree to take their case because their chances of winning are very low, which clearly becomes an obstacle to their access to justice. Another problem is that, in some countries, predicting the outcome of a trial is frowned upon, since it is considered that it may allow litigants to seek the jurisdiction that would be most favorable to them. For example, in France, this type of analysis was banned and it was decided to omit the name of the judge when publishing rulings, to prevent his or her judgment record from being analyzed. But this represents a step backwards in terms of transparency and accountability.
- Finally, in Chile, the Public Defender's Office developed a virtual assistant to support defense attorneys in hearings where it is determined whether or not an accused person will continue his or her trial in liberty. Based on information on the cases presented by the Prosecutor's Office and the possible background of the accused, the tool (which has been "trained" from previous decisions of judges) predicts how the judge will rule based on these facts and proposes jurisprudence to the defender that can be used to develop the defense strategy. This application seeks to help defenders prioritize their cases, based on the probabilities provided by the application. But it also allows them to evaluate their performance, because when they get a result different from the prediction, it seeks to understand the reasons for this result and, if necessary, can lead to conversations with the defenders to discuss ways to improve their work.
These tools, which present advantages and challenges, have often been developed thanks to significant investments by justice institutions -judicial authorities, prosecutors' offices, public defenders' offices-, legal offices or companies that offer their services to these institutions or offices, specifically to address their own problems.
The 'democratization' of AI access
However, a new trend has emerged since the end of 2022, with the emergence of generative AI tools, such as Chat GPT or Bard. These tools propose, at an affordable cost, services that, although not originally developed for this purpose, can be very useful for the procurement and administration of justice and that, in fact, have already begun to be used by justice operators in many parts of the world. This has been explained by Juan Corvalán, founding director of the Innovation and Artificial Intelligence Laboratory of the Law School of the University of Buenos Aires, in a conversation organized last October thanks to the support of the Friedrich Naumann Foundation for Freedom, an institution that has been key to the development of our line of research on digital justice.
Some of these tools, for example, can transcribe and summarize video or audio-recorded hearings; others offer simultaneous interpretation into another language, and some can even summarize and analyze large amounts of documents or suggest arguments about a case.
While it may mean facilitating and streamlining their tasks, the use of these tools by justice operators presents some challenges. First, the inputs with which generative AI tools work must be online, so they become accessible to everyone, which can be a problem if they contain personal data. Another relevant issue concerns the 'hallucinations' - that is, false, wrong or invented data - that these tools can yield. The Brazilian National Council of Justice has already opened an investigation against a judge who cited a case law invented by Chat GPT in one of his sentences.
This demonstrates the importance for justice institutions and law firms to carry out training courses so that their operators know better the potential and limitations of these tools, but also to "overcome their fears", as Laura Márquez, co-founder of the Innovation Center for Access to Justice, and Pablo Pruneda Gross, coordinator of the Law and Artificial Intelligence Research line of the Institute of Legal Research of the UNAM, emphasized during the round table organized by the AMIJ. In this regard, UNESCO has developed training courses for justice operators, as well as a toolkit for judicial authorities that can be very useful.
Nonetheless and, above all, it is urgent that these institutions regulate the use of these tools by their operators, so that they know how to properly use them, while also recognizing what shouldn´t be done. In this regard, the laboratory led by Juan Corvalán, developed Guidelines for the use of text-generative AI and ChatGPT in the Justice system, which can serve as inspiration for interested institutions.
Antidote against corruption
These conversations should not only be of interest to all of us who work in or study the justice system, but also to our governors and legislators. The reason is clear: in the face of the recurring complains by the head of the Executive and many members of the Legislature regarding the problems of corruption that exist in the Judiciary; digital justice, with all its potential to prevent corruption, can be an instrument of defense. Such potential can be expressed through the implementation of platforms to file briefs and/or consult the file remotely, which can reduce contacts between judicial officials and employees to those strictly necessary, contacts that can mean possible bribe requests. On the other hand, digital justice can also reduce the possibilities of file manipulation (since an electronic file cannot be altered or disappear without leaving a traceable digital footprint) and case assignment (as long as a reliable and auditable system is developed).
These conclusions have not only been drawn through our research, but several specialists in the fight against corruption have reached similar outcomes. Such is the case of Sean Hagan, professor at Georgetown University Law School and former IMF Legal Counselor, during a Study Tour on the topic of "Stopping Corruption: Ensuring Government Integrity and Transparency" organized by the Naumann Foundation a few months ago, in which I had the opportunity to participate.
Having a clear understanding of the prior reasons exposed, it is encouraging that actors in the justice system are starting to contemplate digital justice as a priority; at the same time that AMIJ is considering developing a permanent working group on the topic. It would be desirable that, instead of reducing the budget of the Federal Judiciary the Executive and the Legislature push this agenda forward.
 In the Opinion of the Budget and Public Account Commission with the Draft Decree of the Federal Expenditure Budget for fiscal year 2024, one of the arguments to reduce the budget of the Federal Judicial Branch was related to the excessive expenses of the institution in terms of computer equipment and informatics systems.
Laurence Pantin. Coordinator of the Transparency in Justice Program of México Evalúa. Thanks to Pablo García for his review. X: @lpantin.
Read the original text (Spanish) here: https://eljuegodelacorte.nexos.com.mx/la-justicia-digital-un-camino-sin-retorno/