Report 2627
The recent sentence of the First Labor Court of Cartagena that resolved a guardianship action, in second instance, in which inputs from the ChatGPT application were included, opened the debate in the country because it would be the first time that artificial intelligence (AI) is used. to motivate the sense of a failure.
The sentence orders the exemption from the payment of moderating fees and co-payments of the EPS to a minor with autism who does not have financial resources.
There is controversy in the country, voices have been heard for and against the use of this chatbot based on artificial intelligence that solves questions and is capable of delivering several paragraphs with some precision.
El Universal contacted Judge Juan Manuel Padilla García to explain the reasons he had for using this tool when writing part of the motivation for the guardianship sentence. This told us.
What reasons did you have for using the ChatGPT tool to pass the sentence?
Law 2213 of 2022 incorporated the use of ICT in judicial proceedings. It is a binding and imperative mandate. Currently, the use of video calls to conduct remote court hearings is common. Even physical documents were completely replaced by digital PDF files.
The ChatGPT tool is useful for building logical text chains, once the judge defines the legal problem to be resolved and establishes the legally relevant facts. The main advantage that motivated us to take a step towards the use of AI is the optimization of time in drafting court documents.
But we are not only thinking about optimizing times in the elaboration of sentences, in reality we see the glass half full and we intend in the short term through AI to shape procedural measures that allow the judicial process to be promoted, in such a way that the procedures are automated. Car projects such as: set a date, obey the superior, file or issue authentic copies. The creation of document flows is a reality in the private sector, but in the public sector it is also needed.
What is the potential of using ChatGPT or similar tools in the development of the tasks of a judge?
We must recognize with concern that not all judges in the country are trained in the use of new information technologies. The Superior Council of the Judiciary has made a titanic effort to overcome the technological barriers that imply administering justice in times of virtuality. The Rodrigo Lara Bonilla Judicial School through institutional channels has promoted different training for the correct use of ICT. However, we need to move forward and propose institutional goals focused on the concrete use of AI in judicial matters.
I am completely sure that to the extent that we have judges with competencies in technology, in the use of AI, we could solve major problems that afflict justice, including judicial congestion. We require timely justice and I understand that the correct use of AI can facilitate it.
The court ruling ordered the exemption of a minor with autism who does not have financial resources from paying EPS moderator fees and co-payments.
Do you consider that the use of ChatGPT or similar tools can somehow hinder the judge's criteria when making their decisions?
Society should know that judges in Colombia are highly qualified lawyers. They are prepared people, who have come to their position after overcoming great academic challenges. They are even specifically trained to occupy such dignity. The use of ChatGPT rather than hindering the judge, what will really allow you to validate the reasoning of the judicial decision and obtain a chain of texts built with speed and precision that you can use as support in the rationale for the decision.
Can ChatGPT be considered a tool with a sufficient and reliable level of development for the administration of justice?
I understand that the application is in continuous development, so it will not be developed enough to trust 100% of the judicial decision. However, society must ask itself what are the issues that are resolved daily and mostly by judges on a day-to-day basis. To tell the truth, judges in their daily activities, in general terms, resolve repetitive and similar issues. And these are the issues that congest justice. We do not intend to solve the great dilemmas of justice through GPT technology, but we do seek to speed up everyday problems, I mean typical day-to-day actions.
Should the AI tool provide different answers to the judge's criteria, how would this dilemma be resolved?
Human thought will always triumph. The judge is a "sentient being", hence the sentences reflect the critical and pondered feeling of the judges. Judges administer justice for human beings. The prepared judge has sufficient discretion and autonomy to recognize injustice, even if it comes from the analysis of erroneous data on behalf of the AI.
Can we consider that the use of AI will replace at some point the work of the dispatch substantiator?
No. But I do envision a short-term scenario in which the substantiating lawyers of judicial offices, supported by AI, can structure responses to specific legal problems correctly, using the least amount of time possible. The trained AI will have the ability to recognize different variables to a legal problem, finally allowing the judicial operator to adopt the best criterion that qualifies the justice of the specific case.
What do you say to those who have raised concerns that this is an unethical and responsible ruling?
It's easy to criticize. I have noticed that many criticisms come from qualified sectors, but that have never issued a court ruling. It is not the same to dictate a class on guardianship action, than to issue a guardianship sentence. I say it with respect and with criteria of authority, I am a university professor and a judge. However, without the criticism there would be no debate, so I invite you to continue discussing the issue at the national level.
I cannot defend the sentence that gave rise to the discussion by this means, since the sentences are defended in the courts. The application of AI in the revision of judicial texts in no way compromises the ethics of the judge, at the end of the day who is responsible for the decision is the judge.
Does AI have the vocation to change the meaning of a labor failure?
Yes. The AI recognizes, in an indexed manner, valuable information circulating on the internet, which the judge may not be aware of. The value of such a finding can do justice and prevent the judge from committing a judicial error. It is the duty of the judge to corroborate the information in any case. In the future, since the cases are repeated, the validation parameters will be minimal, optimizing the judicial productivity process.
Would you accept that the AI replace the lawyers in the defense of any of the parties in a process before your office?
No. For now it is not possible. But the use of AI on behalf of lawyers will require the judge to be more committed to the judicial decision. It is as if the lawyer arrived at the office with an opinion of his argument, backed by the AI.