Artificial Intelligence Bill
Colombia · Law · Ministry of Science, Technology and Innovation
Our summary
Colombia introduced a unified bill to regulate artificial intelligence development and use according to principles of equity, transparency, ethics, innovation and national sovereignty. The bill applies to AI systems in the country and establishes a National Supervision Authority for AI with powers to set standards, certify systems, assess risks and protect fundamental rights. It also includes provisions for regulatory testing zones, research investment, workforce development and public participation in technology decisions.
Written by the Index in our own words. It is not legal advice; the official text is the authority.
Inside this policy
Every section of this policy in order. Our notes summarise each one in plain language and do not quote the official text, which is the authority.
Drafted by the Index An automated summary of the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
9 sections
Jump to a chapter
1 Institutional governance with technical approach
Creates a National Authority for AI led by MinCiencias, responsible for coordinating public policy, assessing risks, issuing technical guidelines, and operating regulatory sandboxes. Coordination involves relevant ministries and agencies with participation from industry, academia, and civil society for intersectoral and territorially-aware management.
2 Risk classification by level
Categorizes AI systems into four risk tiers: critical systems require prior review without advance prohibition; high-risk systems must undergo impact assessments with transparency and human oversight requirements; limited-risk systems need basic disclosure measures; and low or null risk systems are subject to monitoring and self-regulation. This proportional approach prevents unnecessary barriers to legitimate technology development.
3 Education, training, and research
Integrates AI education throughout the educational system, supports teacher training, and promotes learning in data science and computational thinking. Encourages participation from historically marginalized populations and strengthens academic institutions and regional ecosystems to build AI competencies.
4 Labor protections and workforce transition
Requires that workplace AI use respects human dignity and worker rights, with meaningful human review of significant employment decisions and transparency safeguards. Supports labor adaptation through reskilling programs and training initiatives to help workers transition as automation and technology transform employment.
5 Inclusion, accessibility, and differentiated approach
Applies ethnic, gender, territorial, disability, and sexual orientation perspectives across all measures. Mandates that AI systems prevent discrimination, respect cultural diversity, and function in multiple languages and formats. Promotes technologies suited to Colombian contexts and ensures equitable access to AI benefits.
6 Regulatory sandboxes for supervised innovation
Establishes controlled testing environments particularly for high-risk AI systems, enabling developers to experiment with institutional oversight, ethical evaluation, and rigorous monitoring without anticipatory impediments to responsible innovation.
7 Rights, principles, and actor obligations
Sets guiding principles including respect for fundamental rights, explainability, data quality, transparency, responsible governance, and nondiscrimination. Establishes differentiated obligations for developers, marketers, implementers, and users covering traceability, risk management, and technical supervision throughout the system lifecycle.
8 Accountability, supervision, and control mechanisms
Establishes a supervision regime with graduated corrective measures prioritizing technical guidance over punishment. Contemplates targeted penal code updates addressing only intentional conduct causing grave harm. Existing agencies, rather than new bodies, will conduct oversight.
9 Intellectual property and ethical input traceability
Balances rights between AI technology developers and providers of data, content, or works used for model training. Promotes traceability of training inputs while respecting copyright, image rights, and related protections. Encourages open licensing where feasible and appropriate.
Controls
What this policy requires, forbids, grants or penalises, one control at a time and in our own words. Each control links to the official text, which is the authority. This is not legal advice.
Drafted by the Index Drafted automatically from the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
Themes covered 22 of 35
Choose a theme to see only its controls. About the themes
Risk and scope 4 of 4
Data and privacy 2 of 4
- Data subject rights and consent, not covered
- Cross-border data transfer, not covered
Transparency and records 3 of 4
- Labelling of synthetic content, not covered
Technical safeguards 2 of 4
- Accuracy, robustness and cybersecurity, not covered
- Testing and evaluation, not covered
Assurance and lifecycle 2 of 4
- Registration and notification, not covered
- Incident and breach reporting, not covered
Rights and protected groups 2 of 4
- Biometric identification and surveillance, not covered
- Whistleblowing, complaints and redress, not covered
Roles and organisation 4 of 6
- General purpose and foundation models, not covered
- Public sector use and procurement, not covered
Enforcement and sectors 3 of 5
- Sector rules: health, not covered
- Sector rules: finance and insurance, not covered
The full list of controls could not be loaded, so only the first 27 are shown.
1 3 controls
- ObligationApplies to: Public authorityDrafted by the Index
A public authority should establish a National Authority for AI led by MinCiencias to coordinate public policy, assess risks, issue technical guidelines, and run regulatory sandboxes.
- ObligationApplies to: Public authorityDrafted by the Index
The National Authority for AI should collaborate with relevant ministries and agencies including MinTIC, MinJusticia, SIC and the Procuraduría to coordinate AI governance.
- ObligationApplies to: Public authorityDrafted by the Index
The governance structure should include participation from industry, academia and civil society in AI policy management with territorial awareness.
2 5 controls
- ObligationApplies to: ProviderDrafted by the Index
AI systems must be classified into four risk tiers: critical, high-risk, limited-risk and low or null risk, each with proportionate regulatory obligations.
- ObligationApplies to: ProviderDrafted by the Index
Critical-risk AI systems must undergo prior review without advance prohibition based on potential serious impact on fundamental rights.
Applies when: risk tier: critical
- ObligationApplies to: ProviderDrafted by the Index
High-risk AI systems must undergo impact assessments and implement transparency, traceability, data governance and human oversight measures.
Applies when: risk tier: high
- ObligationApplies to: ProviderDrafted by the Index
Limited-risk AI systems must include basic transparency and traceability measures to appropriately inform users.
Applies when: risk tier: limited
- Applies to: ProviderDrafted by the Index
Low or null risk AI systems should be subject to monitoring and self-regulation rather than extensive regulatory requirements.
Applies when: risk tier: low or null
3 4 controls
- ObligationApplies to: Public authorityDrafted by the Index
Public authorities should integrate AI education across all levels of the educational system and support teacher training in AI-related topics.
- ObligationApplies to: Public authorityDrafted by the Index
Educational programs should promote learning in data science and computational thinking as part of AI capacity building.
- ObligationApplies to: Public authorityDrafted by the Index
Public authorities should support participation of historically marginalized populations in AI education and training initiatives.
- ObligationApplies to: Public authorityDrafted by the Index
Public authorities should strengthen academic and regional ecosystems to develop AI competencies and capacity.
4 4 controls
- ObligationApplies to: DeployerDrafted by the Index
Deployers must ensure that AI use in workplaces respects human dignity and worker rights in all employment contexts.
- ObligationApplies to: DeployerDrafted by the Index
Deployers must apply meaningful human oversight to significant employment decisions made or influenced by AI systems.
- ObligationApplies to: DeployerDrafted by the Index
Deployers must ensure transparency and accountability in the use of AI systems for employment decisions.
- ObligationApplies to: Public authorityDrafted by the Index
Public authorities should implement reskilling programs and training initiatives to help workers adapt to automation and technological change.
5 2 controls
- ObligationApplies to: AnyoneDrafted by the Index
Providers and deployers must design AI systems to prevent discrimination and respect cultural diversity across ethnic, gender, territorial, disability, and sexual orientation dimensions.
- ObligationApplies to: ProviderDrafted by the Index
Providers must make AI systems accessible in multiple languages and formats to serve diverse Colombian populations.
6 1 control
- Applies to: Public authorityDrafted by the Index
Public authorities should establish controlled testing environments for AI systems, particularly high-risk ones, to allow supervised experimentation with ethical evaluation and monitoring.
7 2 controls
- ObligationApplies to: AnyoneDrafted by the Index
All actors involved in AI systems must maintain traceability, conduct risk management, and ensure technical supervision throughout the system lifecycle.
- ObligationApplies to: AnyoneDrafted by the Index
Providers and deployers must respect fundamental rights, ensure explainability, maintain data quality, provide transparency, and apply responsible governance in AI systems.
8 3 controls
- ObligationApplies to: Public authorityDrafted by the Index
Public authorities must establish a graduated supervision regime prioritizing technical guidance and corrective measures over sanctions.
- ObligationApplies to: Public authorityDrafted by the Index
Existing government agencies, not new bodies, shall conduct supervision and oversight of AI systems.
- ObligationApplies to: Public authorityDrafted by the Index
The Penal Code will be amended to criminalize only intentional AI conduct causing serious harm.
9 3 controls
- ObligationApplies to: ProviderDrafted by the Index
Providers and deployers should promote traceability of the data and content used to train AI systems.
- ObligationApplies to: ProviderDrafted by the Index
Developers and providers must respect copyright, image rights and related protections when using data and content for model training.
- Applies to: ProviderDrafted by the Index
Developers and providers should promote and use open licenses for training data and content when feasible and appropriate.
Change history
- Official text wordingSource for this change (opens in a new tab)
- Official text wordingSource for this change (opens in a new tab)
- Official text wordingSource for this change (opens in a new tab)