Based on an official response obtained through a right-to-petition request
This analysis is based on Official Response DEAJTDIFO26-976, issued on 6 July 2026 by Colombia’s Judicial Administration Executive Directorate in response to a formal right-to-petition request.
The response confirms the current institutional framework for AI use within the Colombian Judicial Branch, including disclosure, traceability, human-verification and explainability requirements.
Artificial intelligence is increasingly used to research, summarise and organise legal information. In judicial proceedings, however, the relevant question is no longer simply whether a judge may use AI.
The more important questions are: how was AI used, what was independently verified, and what record must be available to the parties?
In an official response issued on 6 July 2026, Colombia’s Judicial Administration Executive Directorate confirmed that the Colombian Judiciary has binding guidelines governing the use of artificial intelligence by judges, magistrates and judicial employees.
The response refers to Agreement PCSJA24-12243 of 2024, which establishes principles for the respectful, responsible, secure and ethical use of AI within the Judicial Branch.
For companies, in-house legal teams, foreign counsel and litigants involved in Colombian proceedings, this development matters. AI may assist judicial work, but it cannot replace human reasoning, judicial assessment or accountability.
What may AI be used for in Colombian judicial work?
According to the official response, AI tools may support certain tasks, including:
- Searching, collecting and summarising information.
- Suggesting legal sources for consultation.
- Classifying and referencing case law.
- Improving drafting, proofreading and summarising texts.
These uses are subject to human control and verification.
This distinction is important. AI may assist with efficiency and information management, but it does not become the decision-maker.
What may AI not be used for?
The Judicial Administration confirmed that AI should not be used for:
- Assessing evidence.
- Evaluating facts.
- Making value judgments.
- Issuing judicial decisions based exclusively on AI-generated results.
This approach is consistent with Colombia’s Constitutional Court decision T-323 of 2024, which held that AI cannot replace the judge’s human reasoning. Judicial use of AI must respect, at a minimum, principles of non-substitution of human reasoning, transparency, responsibility and privacy.
In practice, a judicial decision must remain the product of the judge’s independent legal and factual assessment.
Human verification is not optional
The official response also clarifies the minimum standard of verification. When AI tools are used, judicial officers must verify:
This is a critical safeguard. Generative AI may produce inaccurate legal citations, outdated rules, incomplete analysis or references that do not exist.
For legal professionals, the practical lesson is clear: AI-assisted research cannot replace source validation. Every authority, legal rule and factual assertion that is material to a case should be independently checked.
What must be disclosed in the record when AI is used?
One of the most significant aspects of the Colombian framework is its traceability requirement.
According to Article 10 of Agreement PCSJA24-12243, the use of AI systems must be expressly reported in the relevant case file, folder, procedural action, document or decision.
This requirement turns transparency into a practical documentation duty.
For parties and counsel, it creates a relevant procedural question: where AI played a material role in a judicial action or document, is there enough information to understand how it was used and whether adequate human review took place?
Why explainability matters
The official response further confirms that the Colombian Judicial Branch contemplates internal explainability standards.
AI tools used by judicial officers should be capable of providing evidence, reasons or justifications regarding their processes and outputs in a way that is understandable to internal and external users. In institutional projects, the Judiciary prioritises tools that provide greater transparency regarding training data and traceable information about inputs, outputs and operation.
For international businesses and legal teams, this is particularly relevant where judicial proceedings involve sensitive business information, confidential documents or personal data.
A technology solution should not become an opaque “black box” that prevents parties from understanding how a material result was generated.
Does improper AI use automatically create professional or disciplinary liability?
Not automatically.
The official response makes an important distinction: Agreement PCSJA24-12243 does not establish a specific disciplinary sanction, nor does it create an autonomous disciplinary or ethical regime for lawyers in private practice.
Any potential disciplinary assessment must be made under the applicable legal framework and by the competent authorities, taking into account the circumstances of the specific case.
That said, the absence of an automatic sanction does not make improper AI use irrelevant. Inaccurate citations, insufficient verification, lack of transparency, disclosure of confidential information or the replacement of human analysis may each create legal, procedural, ethical or reputational risks depending on the matter.
Five safeguards for legal teams using AI in Colombia
Whether preparing litigation documents, internal legal memoranda, contracts or regulatory submissions, legal teams should consider at least the following safeguards:
Verify every legal source
Confirm that each case, statute, regulation and administrative authority exists, remains applicable and supports the proposition for which it is cited.
Preserve human legal judgment
AI output should be treated as preliminary support, not as a substitute for legal analysis tailored to the specific facts, evidence and commercial context.
Protect confidential and personal information
Before entering material into an AI tool, assess whether it contains trade secrets, privileged information, personal data or sensitive commercial information.
Maintain a clear internal record
For sensitive or high-value work, keep records of the tool used, the purpose, the human review performed and the key sources ultimately relied upon.
Establish a responsible-use policy
Organisations should define which tools may be used, for what purposes, what information may not be entered, and which review and approval controls are required.
Key questions for parties involved in Colombian proceedings
Where there are concerns about the use of AI in a judicial proceeding, the following questions may be useful. Tick each one as you review your matter:
- Does the decision, procedural action or document expressly disclose the use of AI?
- Does it identify the tool, model and provider?
- Does it explain the date, purpose and method of use?
- Are prompts or instructions identified where required?
- Is AI-assisted text clearly distinguished?
- Are there legal citations or factual assertions that require independent verification?
- Does the decision reflect independent judicial reasoning?
- Are there indications that evidence or facts were assessed through automated means?
- Could confidential, personal or protected information have been exposed?
- Could insufficient transparency affect due process, the right of defence or the reasoning of the decision?
These questions do not by themselves establish the validity or invalidity of a judicial action. They do help identify when a more rigorous legal and technical review may be required.
AI traceability and legal-risk assessment in Colombia
At Nieto Lawyers, we help companies, legal teams and parties assess the legal implications of AI use in sensitive documents, judicial proceedings and regulated activities in Colombia.
Our review may include:
- AI traceability and disclosure assessment.
- Verification of legal sources and AI-assisted content.
- Due process, reasoning and right-of-defence risk analysis.
- Confidentiality, trade-secret and personal-data risk assessment.
- Internal governance and human-oversight recommendations.
- Practical AI-use protocols for Colombian operations.
Is your legal team using AI in contracts, litigation documents, regulatory submissions or other sensitive materials in Colombia? We would be pleased to assess the applicable risks and help structure a responsible, traceable and legally sound approach.
Request an AI Legal Risk Assessment
We assess AI traceability, transparency, source verification, confidentiality and due-process risks according to the particulars of your matter.
Colombia AI Legal Risk Diagnostic
AI use in a Colombian legal or regulatory context can create risks that are not always visible at the drafting stage: unverifiable legal sources, insufficient human oversight, confidential-information exposure, personal-data issues, and lack of traceability.
Nieto Lawyers offers a focused Colombia AI Legal Risk Diagnostic for companies, in-house legal teams and foreign counsel using AI in contracts, regulatory submissions, internal legal work or Colombian proceedings.
The diagnostic identifies:
- The AI uses that create the highest legal exposure.
- Gaps in source validation, human oversight and traceability.
- Confidentiality, trade-secret and personal-data risks.
- Immediate actions needed to build a legally defensible AI-use framework in Colombia.
Does your organisation use AI to prepare legal documents, manage sensitive information or support decisions affecting Colombia? Request a focused diagnostic to identify the key legal risks before they become a regulatory, contractual or litigation issue.
Request a Colombia AI Legal Risk Diagnostic
Identify the highest-exposure AI uses, oversight and traceability gaps, and the immediate actions to build a legally defensible framework.
Frequently asked questions
Yes. Colombia’s Judicial Branch allows AI to assist with certain tasks, such as research, information synthesis, case-law classification and drafting support, subject to human verification and control.
No. The applicable guidelines indicate that AI should not be used to assess evidence, evaluate facts, make value judgments or issue judicial decisions based exclusively on AI-generated output.
The framework requires AI use to be reported in the relevant case file, action, document or decision, including information about the tool, model, provider, date, purpose, method of use, prompts or instructions and incorporated text.
The cited Agreement does not regulate lawyers’ professional conduct directly. However, lawyers remain responsible for validating sources, protecting confidential information and ensuring that the final document reflects accurate legal analysis.
Not automatically. The official response confirms that the Agreement does not create a specific disciplinary sanction. Any liability analysis depends on the applicable legal regime, the competent authority and the facts of the case.
- Official Response DEAJTDIFO26-976, Judicial Administration Executive Directorate, 6 July 2026.
- Agreement PCSJA24-12243 of 2024.
- Constitutional Court Decision T-323 of 2024.
This article is for general informational purposes and does not constitute legal advice for a specific matter.
