Alexandra Mousavizadeh and Sebastian Hervas-Jones spoke to Victor Munoz, Director of the Administrative Department of the Presidency of Colombia, and Armando Guio, IADB consultant in AI policy. Here is a summary of what was discussed
Colombia has a dynamic and quickly maturing AI ecosystem. It is making impressive progress in our AI Index rankings, and based on these discussions with Victor and Armando, we anticipate that this will continue. Victor’s keen leadership, and Armando’s ambitious plans and commitment to the safe and inclusive development of new technologies appear to be a recipe for success.
Alexandra first asked Victor about his background and how he came to be in this position.
Victor introduced himself and explained that his career began in the private sector, where he created two tech startups, after this he worked as the CIO of a global company with operations in Latin America. In 2018, the Colombian President called him and said that the government wanted to introduce a CIO role in the Colombian government, which would be responsible for the deployment of Colombia’s Digital Strategy, a document which included a strong component in AI.
Victor agreed and initially joined as the Presidential Advisor for Economic Affairs and Digital Transformation. The function of this role was to coordinate and advise the implementation of the digital transformation strategy throughout all entities of the Colombian government, as well as support all sectors within the economy. Since then, Victor’s role has grown to the ‘Director of the Administrative Department of the Presidency of Colombia’ and Colombia has published their “Digital Transformation and AI Policy”. This reflects both Colombia’s keen focus on artificial intelligence and Victor’s expertise.
Alexandra’s next question concerned talent; in many of our discussions with ministers and policy makers, a recurring issue seems to be the development of domestic talent. Therefore, the question of how Colombia is tackling this issue was particularly interesting.
Victor stated that first, it is fundamental to mention that Colombia has the public policy “National Policy on Learning Technologies (CONPES de Tecnologías para Aprender)”. This national policy promotes innovation and development of skills through digital technologies in students of preschool, elementary and secondary education in the official sector, through the following guidelines:
– Ecosystems of Educational Innovation
– Design of curricular orientations
– Training processes
– Creation and dynamization of environments
– Talent development and management
– National Debt Strategy
– Students with STEM skills and innovative schools
In the current situation, Colombia has had to move from a face-to-face model with 9.5 million students to a virtual education model. For this they designed different digital education offers, amongst them ‘Aprender Digital – Contenidos para Todos’, a portal that has 80,000 digital and ‘curricularized’ courses. This platform offers different courses for students of primary, secondary and middle school.
The previous program has been linked with “Computers to Educate (Computadores para Educar)” which seeks to promote access to ICT and present a pedagogical appropriation through the provision of computers to students in different areas across the country. This program aims to give students access to computers, which have preloaded “Aprender Digital”. This will allow them to have full on STEM education through primary, secondary and middle school.
Victor continued, by explaining that the Presidency has promoted several initiatives on digital learning, such as: MissionTIC and 300,000 citizens trained in development, programming and source code. The first program seeks to train 100,000 young and adult Colombians in programming, to face the challenges of the Fourth Industrial Revolution. The second strategy is aimed at strengthening the technical and social-emotional skills of the SENA (national apprenticeship service) apprentices in areas related to programming, web and software design and development, by incorporating SENA’s own technical and technological resources in addition to resources from the information technology sector.
Victor’s team has also implemented the program “Digital citizenship” (Ciudadania Digital). This program intends to impact 5,000,000 citizens in digital skills from all socio economic backgrounds and will be developed through a series of online tools and classes.
The next question Alexandra posed to Victor concerned ethics. She first explained that we are currently building an ethics pillar into our global AI index, which will act as a tool for tracking nations’ success in sufficiently considering the ethical implications of AI development. Therefore both Alexandra and I were curious how specifically Colombia is tackling difficult ethical questions. Alexandra asked whether it is through broad overarching principles, sector specific guidelines, or whether it was negative outcome avoidance focused?
Victor responded by highlighting that the present government has established a clear and long term road map towards the proper adoption of AI in the country. This has been achieved by stipulating 15 strategic main points, in which AI must be implemented thus to create a holistic adoption of emerging technologies in the country.
These points are the following:
1. Creating an AI Market
2. Prioritize market – creating innovations
3. Evidence-based policies and design of impact metrics
4. Regulatory experimentation (Sandbox)
5. Data infrastructure for AI
6. AI Market as a tool to achieve social equality and inclusion
7. Ethical framework for Artificial Intelligence and safety
8. Credible commitments and consensus building
9. Education and digital skills policies based on experimentation
10. Strategic role of academic research in the development of an AI market
11. Attract international talent and skilled labor
12. Evidence-based policies on the future of labor
13. Government as a facilitator and consumer of AI systems
14. Effective implementation of the AI strategy
15. Continuous interaction with the international community and global experts
Under the development of these documents, the Presidential Advisory for Economic Affairs and Digital Transformation, of the Presidency of the Republic of Colombia has created its policy design on the future of AI in Colombia, therefore creating an initial route towards smart regulation and smart parameters, as follows:
1. Ethical framework for the Implementation of AI in Colombia.
2. Conceptual model for the design of regulatory Sandboxes and Beaches in AI.
3. Governance model of data infrastructure for the development of emerging technologies
4. Sandbox on privacy and AI projects
5. Task forces for the development and implementation of the AI in Colombia
6. International Council for AI
Victor explained that it was also fundamental to highlight that this approach has two lines of focus. First is a technical approach to ethics. For this they are implementing a sandbox on privacy by design and by default in Artificial Intelligence, that includes ethics as a driving pillar. The objectives of this sandbox approach are as follows:
- To establish criteria to facilitate compliance with data processing regulations, and indicate procedures for their full implementation.
- To suggest or recommend, if necessary, adjustments, corrections or adaptations of Colombian regulations to technological advances.
- To ensure that the proper processing of personal data is an essential component of the design and implementation of artificial intelligence systems.
- To promote the creation of artificial intelligence systems that, from their design and during their execution, are respectful of the rights of individuals with regards to their personal information and the regulation on the treatment of personal data.
- To support and advise companies on matters of personal data protection, in order to mitigate the risks associated with the implementation of artificial intelligence.
- To consolidate a preventive/proactive approach to human rights protection in AI projects.
The second line of their approach consists in creating regulatory guidelines to support various agencies in preventing bias. Victor explained that the prevention of bias is vitally important in Colombia, as much of the data that has been collected in the past carries with it problematic forms of bias, such as bias around gender equality. To Victor, it seemed essential to develop the systems that will define the future of Colombia in a way that benefits everyone.
Armando Guio Español
I enjoyed speaking to Armando Guio Español. I have a strong feeling that in twenty years from now, when looking back at the development of technology in Latin America and the key players involved, Armando’s name will pop up just about everywhere. While this interview considers Armando’s role in the development of Colombia’s Digital Transformation and AI Policy, he also works as an international consultant with the Latin American Development Bank, as well between many Latin American Governments, such as Peru and Chile.
I began by asking Armando to talk about his work prior to his current role as IADB consultant in AI policy.
Armando explained that he began as a Colombian lawyer, with a focus on topics surrounding privacy and data protection. This focus later turned towards technology policy and legislation. As a lawyer he was curious about the difficult issues the Colombian government was facing in data policies and decisions in data use, this sparked Armando’s more general interest in the development and implementation of public policy.
Armando explained that his time at the Berkman Klein Centre at Harvard University was extremely formative for his current career. During his time there, the team was supporting Angela Merkel and the German Digital Consulate in their development of a National AI Strategy. This allowed him to learn what certain countries were doing and how exactly they were developing their digital strategy. He went on to explain to me that he witnessed the value of collaborating with a range of relevant entities, not just within a nation, but globally. Working with representatives from Canada, and the UK (who later became the office for AI for the British Prime Minister) as well as the team at Harvard, made for some fast paced and fruitful collaboration, which, at the time, enabled them to assist in the production of a comprehensive AI strategy for the German Government.
It was during this time at Harvard that Armando came across Victor Muñoz. Victor had come to Harvard for a conference, in which they discussed Colombias’ interest in generating and adopting a digital transformation policy. Here Armando suggested to Victor that Colombia should also consider the possibilities and implications presented by the rapid global development of artificial intelligence. Victor said that Colombia was interested in doing this, but at the time, the Colombian team didn’t have the expertise required to produce such a document in AI. However, Victor said to Armando ‘if you are the one suggesting Colombia adopt an AI strategy, then you should be the one to do it.’
Unsurprisingly, Armando gladly accepted this offer and joined the Colombian government to work for the presidency, supporting the National Planning Department, and Victor’s team.
In my next question to Armando, I asked if he could run me through the process of writing and publishing Colombia’s Digital Transformation and AI Policy.
Armando began by laying out the specific steps involved in producing their national AI strategy. The first step being the initial discussions and meetings which laid out the precise focus of their strategy. Armando insisted to the team that they specifically considered AI, not only digital transformation, which was why they entitled the strategy “Colombia’s Digital Transformation and AI Policy”.
With this in mind, Armando said that gathering a diverse and skilled range of relevant entities was vital in producing as comprehensive a strategy as possible.
The next step was to set out specific guidelines and priorities: What do we want our national AI Strategy to focus on; what are the most enticing possibilities presented by AI that we can encourage; and equally, what are the most daunting implications of these technologies that we need to create policies to protect our citizens? Once these questions were answered, Colombia was able to set these general guidelines, or ‘sections of the strategy.’ For example, a government may want to develop (1) a digital skills program that encourages the growth of a strong talent pool and (2) an ethical framework that ensures AI benefits all its citizens whilst causing no harm .
From this step, the team set to work on more specific policies, setting guidelines and establishing working groups for carrying out the actions described in the policy. For Armando and Victor in Colombia, it was essential that the Presidency of the Republic be responsible for developing the ethical framework. Coordination and clear deadlines were valuable, and participation of diverse entities ranging from the president to affected businesses seemed like the only way to ensure that the strategy was applicable and practical.
The conversation surrounding the ethics of artificial intelligence is one often had, yet one extremely hard to legislate and track. Therefore the final question posed to Armando sought to ascertain Colombia’s specific approach to developing regulation to enforce these ethical guidelines.
Armando explained that with the support of the Latin American Development Bank, Colombia had just completed an ‘Ethical Framework for Artificial Intelligence’, which turned out to be the first one in Latin America. Armando explained that the framework is a “soft law” guide for the implementation of ethical and responsible AI in the public sector. The framework took into account the ethical principles from the OECD, UNESCO and from a research conducted by the Berkman Klein Center at Harvard University, and adapted them to the Colombian context. Armando stated that the framework seeks to implement ethical AI in practice, beyond principles. To do this, it establishes a series of ‘measures’ or ‘tools’ to achieve this purpose; such as ‘algorithm ethical assessments’ and ‘codes of conduct and legitimacy evaluation’, among others.
It is also interesting to note that the Colombian government is now considering the implementation of an ethical algorithm register. Armando concluded by explaining that now, the challenge will be to implement these principles and mechanisms to achieve “an ethical AI in Colombia.” He said that he expects to achieve this from his new position as a consultant at the Inter-American Development Bank. This position, Armando said, will also assist in gaining valuable insights that can be taken to other countries in Latin America and the world.