Report
AI already makes decisions in the Colombian state. The rules do not, yet
Key takeaways
- Colombian public AI is real, longstanding, and concentrated in high-risk functions: justice, enforcement, migration.
- No documented system operates under binding requirements for risk assessment, public registration, or measured oversight.
- The ILIA paradox — governance 73.2, adoption 35.0 — is explained by three deficits: institutions, talent, and controls.
- The usable strength exists: Colombia's standout open-data record is the natural foundation for a public register of systems.
- The Cartagena case on generative AI in a court ruling is under Constitutional Court review: the first enforceable rules may arrive through the courts before the policy does.
Colombia designed AI policy earlier than almost anyone in the region — and adopted AI systems earlier than almost any of its rules. Here is what we found building the first public baseline, and what can be done without waiting for a law.
The full document is currently published in Spanish; an English edition is prepared when demand warrants it. The Spanish record carries the complete summary and contents.