LEGAL TECH

AI Courts Where You Vacation

Costa Rica amazes day in and day out with its crystal-clear waters, year-round warm temperatures and pristine rainforest. This Central American country has only 0.03% of the world’s surface, but astonishingly houses 6% of the world’s biodiversity. Due to its natural resources, this nation has become a global pioneer in sustainable travel and ecotourism. But in addition to its natural reputation, it seems that Costa Rica is willing to show the world that they can also lead by example in the application of technology. Recently, the Costa Rican judiciary announced their first AI project to be used in its courts.

By Gabriel Obando-Chacon

February 27, 2021

Luis Quintanilla

How can AI be used in courts?

 The first thing that comes into mind when someone talks about AI in courts is possibly a robot judge. Nevertheless, while this is not entirely impossible, the state of the art is just not quite there yet. The current AI systems are more likely to assist judges and court officials, as opposed to replacing their roles. The present-day technology is still incapable of rendering complex legal decisions on its own (or at least to the standard desired for judicial decision-making) and will likely continue to be designed to assist judicial courts in the near future.

Some countries have already experienced different assistive applications of AI in their judicial systems. For example, it’s likely that you have already heard of the heavily criticized COMPAS, an algorithmic system that is currently in use in several US courts. This AI system is supposed to measure the chances of a prosecuted felon of committing another crime (“recidivism”). This AI system is considered to be assistive as the judge weighs in how heavily she wants to rely on the algorithmic output for her final decision. The algorithm becomes another piece of the decision-making puzzle.

But other countries seem to be convinced that AI could potentially replace judges. In 2018, Estonia announced that they would launch a small claims court whose decisions would be entirely produced by an algorithm, without the assistance of a human judge. However, two years after the project was announced, it has not really taken off. It seems that the task is far too expensive or complex, even for Estonia, a forerunner in the application of technology in tasks related to public administration.

So, what is the Costa Rican judiciary going to do?

The Costa Rican judiciary is keeping things simple. Instead of being overly ambitious, the Costa Rican judiciary have trained an algorithm that is able to categorize the type of requests that a party to a collection proceeding makes before the court. In short, a party files a letter, for example, requesting the court to change their email address to receive court notices, then, the system labels this brief as a “change in notices”, allowing the judge to preemptively know what the brief is about.

While this might not sound revolutionary (and it’s not!) the positive impact that a minor change like this in terms of time consumption could be huge. In a country of only 5 million people, there are over 733.169 active collection cases, assigned to a taskforce of 86 judges nationwide. In other words, that is over 8.500 collection cases per judge. With the AI system implemented, the judges will be able to prioritize their attention to certain briefs as they consult a file. In general terms, involving AI could make a difference in case management. Of course, it will not be the ultimate solution for all issues related with judicial administration, but it’s a start.

What are some challenges in the near future?

Innovating in the public sector is not easy. The Costa Rican judiciary faces many challenges ahead with its first AI project. For instance, the training period for the algorithm will have to be heavily supervised. There is little margin for error in judicial environments and one mistake could have negative effects in thousands of files and people’s lives. Adequately, the Costa Rican judiciary is taking precautions by implementing a pilot project in Perez Zeledon (a town South of the capital) to evaluate the success of the AI system, before it is used anywhere else.

Furthermore, projects like this should catalyze the discussion for further privacy regulations in Costa Rica. As the application of technology becomes the norm in public administration and, in judicial decision-making particularly, clear and enforceable rules will be required to make sure that the citizenry’s data is processed in accordance with the rule of law. As of now, Costa Rica has a privacy act, but this is far from the standards promoted by the EU’s GDPR, which appears to be the metric for this type of regulation in the world. Modernizing the legal framework is not easy, but it is certainly of essence.

Despite these challenges, this small Central American nation appears to be willing to take risks through a technological leap in its courts. Positive experiences with AI in courts have already been reported in Mexico, Colombia and Argentina, so Costa Rica should not be an exception.While the application of technology should always be welcomed, as a society, we should remain vigilant to guarantee that the rule of law and our data are always well protected. 

Gabriel holds an LLM from McGill University. His research focused on privacy concerns related to the collection of data by governmental agencies to implement judicial AI. Before that, he obtained an LL.L from the University of Costa Rica with a concentration in Human Rights Law. He has worked in 'Big Law' and boutique firms in Costa Rica. Most recently, he joined Nelson Champagne Avocats in Montreal where he's part of the class actions team. He's also the co-founder of algoritLaw, a blog that explores the intersection of AI and Law. He's fluent in English and Spanish, and proficient in French.