Colombia Moves to Regulate AI, Balancing Innovation and Risk

From Asimov’s fictional laws to real-world legislation in the Colombian Congress: who is truly in control when algorithms begin making decisions for us? Colombia is actively working to establish a legal framework for artificial intelligence (AI), aiming to foster innovation while mitigating potential risks. This move reflects a global shift towards AI governance, recognizing that voluntary ethics alone are insufficient to ensure responsible development and deployment.

The Need for AI Regulation: Beyond Ethical Pillars

AI is rapidly advancing, demonstrating capabilities in areas like disease diagnosis, text generation, city optimization, and automated decision-making. However, this progress raises critical questions about fairness, impartiality, and safety. While current AI development relies on voluntary ethical principles – transparency, fairness, and explainability – international experience has shown these are often inadequate. Instances of algorithmic bias, misinformation, and data misuse highlight the need for stronger safeguards to maintain public trust. Even leaders in the field, such as OpenAI CEO Sam Altman, acknowledge the risks of unchecked innovation, calling for urgent regulation.

Colombia’s Foundation for AI Governance

Colombia isn’t approaching AI regulation without prior groundwork. The country has already established several key foundations:

• Ethical Framework for AI (2021): A foundational document outlining ethical considerations for AI development.
• National Policy on Artificial Intelligence – CONPES 4144 (2023): A national policy providing strategic direction for AI initiatives.
• Data Protection Authority Guidelines on automated systems (2024): Guidelines from the Data Protection Authority addressing the use of automated systems.

Current Legislative Efforts: Six Bills in Congress

The core of the regulatory discussion is now taking place in the Colombian Congress, with six bills under consideration. These bills address four key areas:

1. Comprehensive Risk-Based Framework: A proposal to classify AI systems based on their risk level, requiring impact assessments for high-risk applications and establishing a specialized national authority. This framework is inspired by international standards, including those from the OECD, UNESCO, and the European Union.
2. Sector-Specific Regulations: Bills focusing on specific applications of AI, such as citizen services and consular operations, emphasizing human oversight and data protection.
3. Protection of Vulnerable Populations: Specific rules governing AI use involving children and adolescents, with a focus on psychosocial impact and digital equity. These rules would ban AI systems that seek to control the physical or mental will of any human being, and those that engage in discrimination.
4. Institutional Strengthening: The creation of a permanent legal commission within Congress dedicated to Artificial Intelligence.

The Path Forward: ‘Smart Regulation’ for Sustainable AI

Colombia aims to regulate AI in a way that doesn’t hinder innovation or fragment the legal system. Effective regulation requires a coherent architecture that is proportional to risk and technically informed. The goal is to create “smart regulation” that fosters public trust and ensures the long-term sustainability of AI technologies. As highlighted by the Ruta N Thought Center, Colombia has the opportunity to build a framework that protects its citizens while simultaneously driving innovation. The key question is whether this can be achieved effectively and in a timely manner.

The proposed bill, once approved, will apply to all individuals and organizations – public or private – involved in the creation, development, implementation, commercialization, import, distribution, or use of AI systems within Colombian territory, as well as those subject to Colombian law even if located abroad. These entities are referred to as “AI Responsible Parties.”