Colombia has published the "Ethical Guide for the Implementation, Development and Use of Artificial Intelligence Systems Artificial Intelligence in Public Entities." This is a document that lays the foundations for future binding legal regulation and sets as a priority the protection of human rights, the dignity of individuals and democratic values in the State's use of these technologies.
Guiding the entire lifecycle of AI systems
The guide, aimed at national public entities, seeks to provide guidance throughout the entire lifecycle of AI systems. From their design and development to their implementation and use, to ensure that their adoption contributes to the common good, strengthens institutional trust and avoids practices of invasive control or automation without human oversight. The document is presented as a strategic instrument to turn AI into a driver of public value for the country, in line with OECD (2025) recommendations.
The text adopts a comprehensive ethical vision based on global principles and international best practices, adapted to the Colombian context and its institutional, social and legal needs. Within this framework, it identifies the fundamental ethical principles that should guide the use of AI in the public sector, the key enablers for successful implementation and the safeguards needed to identify, prevent and mitigate the risks associated with these technologies.
AI as a strategic tool for the State
Among its general conclusions, the guide stresses that AI represents a key strategic tool for modernizing and strengthening the Colombian public sector, with the potential to improve productivity, citizen services and trust in institutions. However, it warns that its implementation must be approached from a comprehensive vision that combines ethical principles, governance frameworks adapted to the national context and sustained investment in critical capabilities such as data, technological infrastructure, human talent and financing.
Alfredo Collosa publishes a book on AI in the public sector
The document insists that, in this vision, the protection of human rights and the dignity of individuals must always prevail over criteria of efficiency or economic growth. One of the central objectives is for public entities to design AI systems aimed at enhancing human capabilities and optimizing processes, avoiding their use as tools of instrumentalization, excessive surveillance or invasive control.
Likewise, it highlights that the success of responsible AI in the government sphere will depend to a large extent on multisectoral collaboration between the State, the private sector, academia and civil society, as well as on promoting GovTech initiatives and public-private partnerships. To this is added the need to consolidate a culture of continuous learning, supported by robust monitoring and evaluation mechanisms that allow strategies to be adjusted as technology and the environment evolve.
Recommendations
The guide includes a detailed set of recommendations aimed at ensuring that the adoption of artificial intelligence in the Colombian public sector is ethical, equitable and generates citizen trust. Among them, it highlights strategic investment in key enablers, starting with data, through the creation and maintenance of a high-quality, representative and accessible public data infrastructure, with guarantees of privacy and security.
On talent, the document proposes implementing mass training programs in artificial intelligence and digital literacy for all levels of public administration, from general users to highly specialized profiles. On infrastructure, it recommends evaluating and adopting computational power solutions, including cloud or hybrid architectures, that respond to AI needs, incorporating criteria of efficiency and environmental sustainability.
Verifying the integrity of information
Another central recommendation is the implementation of comprehensive audit and evaluation systems, through the development of internal and external capabilities to conduct Algorithmic Impact Assessments (AIA) and sociotechnical audits on a regular and mandatory basis, especially in systems considered high-risk. These assessments must include robust mechanisms to verify the integrity of information and proactively mitigate the risks arising from the use of synthetic content or deepfakes. The guide also stresses the need to ensure accountability, identify risks early and establish clear mechanisms for recourse and appeal against AI-assisted decisions.
On transparency and explainability, the document proposes the creation of a public and accessible registry of all artificial intelligence systems used by national entities, detailing their purpose, the data employed, the operating mechanisms and the impact assessments carried out. Likewise, priority is given to the design of interpretable systems, capable of explaining their decisions in clear and understandable language for citizens, especially in sensitive areas such as justice.
Data anonymization
In the Colombian case, the guide states that this registry must be mandatory and that, in compliance with Ruling T-067 of 2025, algorithmic transparency guidelines must include proactive evaluation of the publication of source code, in whole or in part, in high-risk systems, guaranteeing data anonymization and the separation of access credentials to sensitive databases.
Finally, the guide recommends leveraging and scaling successful experiences through the systematic documentation of lessons learned in national and territorial initiatives. These experiences should serve as a basis for replicating and adapting AI solutions in other contexts and government functions, fostering the exchange of knowledge and best practices between entities, with permanent collaboration structures that allow maximizing the impact of public investments in artificial intelligence.
With this publication, Colombia advances in building an ethical and operational framework that lays the foundations for future binding legal regulation of responsible artificial intelligence, with the goal of moving toward a more efficient, transparent and people-centered government.





