By: Judicial Desk
Updated: July 18, 2026
In a landmark decision that bridges the gap between rapid technological evolution and the rigid requirements of criminal law, Colombia’s Constitutional Court has issued a definitive ruling on the use of Artificial Intelligence (AI) in criminal activity. Through Judgment C-212 of 2026, the high tribunal has upheld the constitutionality of a sentencing enhancement for the crime of personal impersonation—commonly known as identity theft—when committed using AI-powered tools.
This decision marks a pivotal moment for the Colombian justice system, confirming that even in the absence of a singular, statutory definition of Artificial Intelligence, the judiciary possesses the tools necessary to interpret and apply criminal statutes to modern, high-tech offenses.
The Core Conflict: Can a Law Exist Without a Statutory Definition?
The controversy surrounding the constitutionality of the enhancement arose from the passage of Law 2502 of 2025, which amended Article 296 of the Colombian Penal Code. The amendment introduced a specific aggravating factor for those who utilize AI to deceive or impersonate individuals.
Petitioners challenging the law argued that it violated the principle of "strict legality" (nullum crimen, nulla poena sine lege). They contended that because the Colombian Congress had failed to provide a precise, codified definition of "Artificial Intelligence," the law was unconstitutionally vague. The argument suggested that such ambiguity granted prosecutors and judges excessive discretion, potentially leading to arbitrary interpretations where any digital tool—no matter how simple—could be misclassified as "Artificial Intelligence," thereby unfairly increasing the severity of a defendant’s sentence.
The Court, led by the ponencia of Magistrate Jorge Enrique Ibáñez, rejected these claims. The justices concluded that the perceived "legislative vacuum" was not a barrier to justice, but rather a flexible reality that requires a functional, rather than rigid, interpretation.
Chronology of the Legal Evolution
The journey toward this ruling reflects the accelerating pace of both digital innovation and regulatory adaptation in Colombia:
- 2019–2022 (The Foundation): The Colombian government issues the first major policy frameworks regarding AI, specifically CONPES 3975 of 2019 (National Policy on Digital Transformation and AI) and Decree 1263 of 2022, establishing initial ethical guidelines for AI adoption in the public sector.
- 2025 (The Legislative Action): Recognizing the rising tide of deepfakes and automated identity theft, Congress passes Law 2502 of 2025, explicitly adding an aggravating factor for impersonation crimes involving AI.
- Early 2026 (The Challenge): Legal scholars and citizen litigants file a constitutional challenge against the law, arguing that the term "Artificial Intelligence" is too vague to satisfy constitutional due process requirements.
- July 18, 2026 (The Ruling): The Constitutional Court publishes Judgment C-212, declaring the law constitutional and providing a framework for how judges should identify AI in criminal proceedings.
Defining the "Indefinable": A Functional Approach
One of the most significant aspects of the Court’s ruling is its refusal to trap itself in a static definition. The magistrates argued that if the law were to define Artificial Intelligence with the scientific precision of a computer science textbook, it would be rendered obsolete within months, if not weeks, by the exponential nature of technological progress.
Instead, the Court adopted a functional interpretation. By synthesizing guidance from international bodies—including the OECD, the UNESCO Recommendation on the Ethics of Artificial Intelligence, and the European Union’s AI Act—the Court identified three core characteristics that define AI for the purpose of the criminal statute:
- Inference from Data: The ability of a system to identify patterns or derive information from large datasets rather than being explicitly programmed for every specific output.
- Autonomous Content Generation: The capacity to produce predictions, recommendations, text, audio, or visual content (such as deepfakes) that would typically require human cognitive abilities.
- Adaptability: A degree of system autonomy that allows it to adjust its behavior or performance based on new inputs or environmental feedback.
"The aggravating factor does not rely on a static dictionary definition," the Court noted. "It relies on the functional reality that the tool performs tasks—previously reserved for human cognition—to facilitate the crime of identity theft."
Implications for the Criminal Justice System
The Court’s ruling establishes a "case-by-case" mandate for judges. The judiciary is instructed to look at the specific technology used at the time of the offense. If a perpetrator uses a sophisticated generative model to clone a person’s voice or likeness to commit fraud, the aggravated penalty applies. Conversely, the ruling explicitly protects the average citizen from over-broad applications; the simple use of basic software, filters, or non-AI digital tools does not trigger the enhancement.
Safeguarding Against Arbitrary Prosecution
The Court emphasized that this ruling does not replace the role of Congress. If the legislative branch eventually decides to draft a comprehensive AI law with a formal legal definition, it is fully within its power to do so. However, until such time, the Court has provided the judiciary with the necessary legal grounding to ensure that the law keeps pace with criminal innovation.
A Warning to Cybercriminals
For those engaged in identity theft, this ruling sends a clear message: the state is modernizing its enforcement mechanisms. By classifying AI-facilitated impersonation as a distinct, aggravated offense, Colombia joins a select group of nations attempting to proactively discourage the weaponization of synthetic media.
Expert Perspectives and Societal Impact
Legal analysts have praised the ruling for its "technological neutrality." By focusing on what AI does rather than how it is constructed, the Court has shielded the law from obsolescence.
"The Court correctly identified that the danger of AI in criminal contexts is the scale and realism of the deception," says an expert in digital law. "By linking the aggravation to the facilitation of the crime through autonomous or inferential processes, the Court has created a robust barrier that still respects the principle of due process."
However, the burden now shifts to the prosecutors. To secure an aggravated conviction, the Prosecutor’s Office (Fiscalía) must now provide technical evidence showing that the tool in question meets the functional criteria set out by the Court. This will likely necessitate a higher degree of cooperation between law enforcement and digital forensics experts.
Conclusion: A Living Constitutionalism
The C3-212 ruling serves as a prime example of "living constitutionalism," where the high court interprets the constitution not as a static document, but as a framework that must adapt to the challenges of the 21st century.
As AI continues to blur the lines between reality and simulation, the Colombian legal system has signaled that it will not stand idly by. While technology may move at an exponential pace, the reach of the law—guided by the wisdom of the Constitutional Court—is now better equipped to ensure that the digital world remains subject to the same standards of integrity and accountability as the physical one.
For now, the mandate is clear: the law is not a frozen relic, but a dynamic instrument capable of addressing the complex intersections of code, crime, and human identity. The era of AI-facilitated impunity is effectively over, replaced by a nuanced, evidence-based legal standard that prioritizes both technological progress and public safety.
For more information on legal developments and judicial rulings in Colombia, subscribe to the Justice Desk newsletter or follow our live updates on our official social media channels.
