Prepared by Jorge Castaño
Blanca is a single mother of a child diagnosed with Autism Spectrum Disorder (ASD), without stable employment and living in extreme poverty (classified as A4 in the Sisbén system). Nevertheless, the EPS (Health Promotion Entity) did not exempt her child from copayments and moderating fees. Furthermore, the EPS denied them the resources needed to cover intercity transportation for medical appointments. Blanca file a legal claim (acción de tutela) and both the first and second instance judges upheld the fundamental rights of the child.
This ruling would have been like many others issued daily by judges in Colombia, but the second instance judge noted in his decision that, to reinforce his reasoning, he consulted ChatGPT 3.5. He asked specific questions regarding the right to exemption from copayments for minors with disabilities, the admissibility of the tutela action, and judicial precedents in Colombia. ChatGPT provided answers, which the judge transcribed into his decision.
The Constitutional Court decided to review the case, as the use of artificial intelligence in judicial decisions raised an urgent national debate regarding due process, judicial ethics, and the role of AI in the administration of justice. That is, the Court needed to determine whether the judgment had been issued by a judge of the Republic or, in effect, by ChatGPT 3.5.
Guidelines for the Use of AI in the Administration of Justice
Based on the above, the Constitutional Court, to prevent violations of due process, established guidelines for the use of AI in the administration of justice, which were subsequently adopted by the Superior Council of the Judiciary. The following are the most significant rules:
- Transparency, the obligation to clearly and precisely disclose the use, scope, and placement within proceedings or decisions obtained using such tools, enabling users and interested parties to fully understand and effectively having the information to challenge them.
- Responsibility, the obligation for users of AI tools to be properly trained and understand the impacts of these technologies, being accountable for the origin, appropriateness, and necessity of AI use, with verification of the information provided.
- Privacy, the duty to safeguard and protect the confidentiality of personal and sensitive data entrusted to the judiciary for the fulfillment of its purposes.
- No Replacement of Human Rationality, ethical and legal impossibility of replacing human action and responsibility in judicial proceedings and decisions.
- Equality and Fairness, eradication of all forms of discrimination related to biases introduced using such technologies and their negative impact on human rights.
- Human Oversight, ensuring effective scrutiny of proceedings and decisions involving AI tools, with access to necessary information and the use of remedies to be resolved by human authorities.
Additionally, Agreement PCSJA24-12243 of December 16, 2024, issued by the Superior Council of the Judiciary, sets forth the criteria for the use of AI by judiciary officials. It permits administrative and support uses, such as assisted drafting, document organization, and case law research, but prohibits the use of AI for evidence evaluation or decision-making. Specifically, the Agreement states that AI systems must not, under any circumstances, replace activities such as the reasoning behind decisions, fact assessment, evidence analysis, legal application, interpretation, and decision-making.
Guidelines for the Use of AI by Lawyers – International
In both the United States and England, various associations have reviewed the fundamental aspects to be considered in the use of AI by consulting and litigating attorneys. The main aspects identified are:
- Reasonably understanding the functioning, capabilities, and limitations of AI.
- Maintaining confidentiality and, if necessary, obtaining client consent for the use of AI tools.
- Determining how to communicate to the client the use of AI when it is related to important decisions.
- Reviewing AI-generated outputs to avoid the presentation of false or erroneous information.
- Establishing institutional guidelines for the use of AI by firm members.
- Finally, these guidelines establish principles to determine when the costs associated with Generative AI (GAI) may be charged to the client, emphasizing that fees must be reasonable and reflect actual costs, without generating hidden profits for the lawyer.
Guidelines for the Use of AI by Lawyers – Colombia
The question that arises is whether there are guidelines for the use of AI by lawyers in Colombia. A general review of Law 1123 of 2007 (Lawyer’s Disciplinary Code) shows that many of the above-listed principles are already contemplated in the Code and, moreover, have the biding effect of legal obligations.
International Guidelines | Correspondence with Colombian Lawyers Disciplinary Code |
Competence (Duty to be competent and updated, including technological competence) | Article 28: Duty to update the knowledge inherent to the practice of the profession. This must include the use of technology. |
Confidentiality (Protecting client information) | Article 28: Duty to maintain professional secrecy, even after services are concluded. Article 28: Duty to act with loyalty and honesty in professional relations. This may involve disclosure and acceptance by the client regarding the use of AI. |
Supervision of Subordinates (Duty to supervise employees and third parties) | Article 28: Extends to supervision to substitute, subordinate, and associated lawyers in contracts, but does not explicitly regulate the supervision of technologies or virtual assistants, unlike the ABA. |
Truthfulness before Courts (Transparency before courts) | Articles 32 and 33: Prohibition against malicious statements, inaccurate citations, or use of false evidence. |
Reasonable Fee Charging | Articles 28 and 35: Duty to set equitable, justified, and proportional fees; prohibition of disproportionate charges. |
Furthermore, considering that judicial decisions must be made by human beings vested with competence and jurisdiction, and that this faculty cannot be delegated to another human being or to an AI system, it must be considered that only licensed attorneys with a professional license are authorized to represent a party in a judicial process and to provide legal advice. Thus, this function cannot be delegated or attributed to an AI system.
Conclusions and Recommendations
- Vigilance over the use of AI by judicial officials is necessary, as improper use could violate the fundamental right to due process, resulting in annulments.
- Law firms should establish guidelines for the use of AI in the provision of services, incorporating international standards and observing the Lawyer’s Disciplinary Code.
- Vigilance over the use of AI by lawyers in judicial representation is also necessary, since improper use may raise objections as to whether judicial representation was carried out by a licensed attorney or by AI, thereby creating annulment risks.
Do not hesitate to contact Brick Abogados if you have any questions or if you would like further details on the topic discussed above.
****
This document is for informational purposes only and does not constitute legal advice, nor does it engage the responsibility or professional opinion of Brick Abogados.