Q&A: Armando Guio Español on AI Ethics and Policy
BKC affiliate drafted Ethical Framework for Artificial Intelligence in Colombia
Increasingly, the public and private sectors alike are working on guidance and principles for artificial intelligence (AI) development, deployment, and regulation. The Colombian government, for example, recently published a draft AI ethics framework, “Ethical Framework for Artificial Intelligence in Colombia.” AI policy expert Armando Guio Español, an alumnus of Harvard Law School, an affiliate of the Berkman Klein Center (BKC), and consultant for the Development Bank of Latin America — drafted the framework for the Colombian government.
On November 24, Armando participated in an expert roundtable discussion on Colombia’s AI ethical framework and its implications for young people. The BKC Policy Practice on Artificial Intelligence hosted the roundtable in collaboration with the Colombian government. BKC Policy Practice: AI is a public interest-oriented program, led by BKC Executive Director Urs Gasser, that helps governmental, nonprofit, and private sector organizations implement AI best practices and turn AI principles into operational realities.
We spoke with Armando about Colombia’s Ethical Framework for Artificial Intelligence, the recent roundtable discussion focusing on youth, and how working at BKC informed his experience creating the framework. Responses were lightly edited for clarity.
You authored Colombia’s Ethical Framework for Artificial Intelligence, which creates a foundation by which Colombia’s public and private sectors can embrace AI while respecting core values such as protecting human rights and privacy. Could you briefly summarize the main points?
The Ethical Framework for AI describes ten principles for designing, developing, and deploying trustworthy AI systems in Colombia. Some of these principles are transparency, explainability, privacy, human control, social benefit, and the prevalence of children’s and adolescents’ rights, amongst others. Each of these ethical principles is considered from a tri-dimensional perspective: data, algorithms, and practices. When we have discussions about transparency and AI, we have different considerations when we talk about being transparent: on the use of data, on the modeling of an algorithm, or in the practices we undergo to implement this technology. Likewise, the framework also carefully considers the algorithmic chain and how principles will be applied in the design, development, or implementation of this technology. Creating the framework has been quite interesting because it has allowed us to discuss the ethical consequences of implementing this emerging technology in Colombia, pointing out that ethics are not only a matter of designing this innovation.
Additionally, the framework serves as an ethical toolkit. It provides a series of options to operationalize these principles in entities from the public and private sectors. We have proposed different methodologies such as algorithm assessment, algorithm auditing, data cleansing techniques, ethical training in coding courses, legitimacy evaluation for AI projects, and the implementation of Codes of Ethics, amongst others. We present a range of options that all entities can adopt per their convenience and own internal procedures.
Finally, the ethical framework sets a series of recommendations for the Colombian government to develop these ethical principles further. The main proposal was for the government to use this framework as a first draft to mobilize the discussion we must have as a society about using such disruptive technology.
What was the process for creating such a framework? What factored into your thinking?
The ethical framework is part of a bigger effort I have been working towards: implementing Colombia’s AI Strategy. In 2019, for example, I led the design process of Colombia’s AI strategy with Victor Muñoz, the Presidential Advisor for Digital Transformation in Colombia. This opportunity allowed Colombia to create one of the first AI Strategies in Latin-America. We defined a series of guidelines, one of them being an ethical framework. Therefore, having this framework is just one step towards properly implementing this policy that Colombia has adopted.
With the support of the Development Bank of Latin America (CAF), I started the task and, I must say, the challenging experience of preparing this first draft. For me, it was essential to look at the existing academic work that supported many of the principles being proposed around the world. I will highlight that the work and literature developed by BKC, especially from the Ethics and Governance of AI Initiative, was essential for this purpose. Additionally, I looked into the evidence that was developed in the last few years about the possible risks arising from the use of this technology. Case studies around the world showed the risks of using AI without proper ethical considerations. The study published by BKC about AI and Human Rights was exceedingly relevant to this point.
Once I was convinced about the importance of having ethical principles, I started thinking about those principles that could become the most relevant for Colombia and our technology ecosystem. I tried to think through this from different perspectives so it could become a national framework. My experience working in different sectors allowed me to consider the interests and needs of other entities. From academia to the private sector and public entities, I wanted these principles to have a horizontal view. I have had, and still have many discussions, but this was a helpful exercise to work on a document that was convincing enough for different stakeholders. I also considered that such an effort required proper research and a quality document that allowed us to have a high-standard discussion. That was one of my main motivations.
You joined the BKC community as a fellow in 2018 and remain active as an affiliate. How did your time at BKC support your work on AI and ethics, and specifically your thinking about AI and youth?
Without a doubt, I would not have been able to work on this project without being part of the BKC community. When I arrived at BKC in 2018, AI was becoming a very relevant issue, but still, in countries such as Colombia, the topic was not seen as a priority. Additionally, there were many myths about this technology and its real potential. My time at BKC allowed me to understand the real features of AI, its characteristics, and how it could impact Colombia and the region. Sharing time with such an exceptional group of individuals through fellows hours, coffee breaks with colleagues from all around the world, Urs’ office and “un-office hours,” were essential for learning, questioning myself, and going beyond my initial thoughts. There were many challenging conversations, and that is only part of the Fellow experience. At the same time, I was able to discuss with many of the authors of some of the papers and research that I was reading at the time, and who were impacting my approach to these topics. That was a great privilege.
I also had the opportunity to work in the Youth and Media office, which was very helpful for the work I was going to develop. Not only were Sandra Cortesi and Alexa Hasse very welcoming, but they also shared with me their research ideas, their perspectives about these topics, and the interesting and innovative things that I could do. I had never considered how important it was to design spaces to discuss with children about AI systems and how to get the most from their approach to this topic. Sandra’s advice also inspired me to explore my creativity, appreciate a multidisciplinary approach, and consider the relevance of design features in the produced documents. I also shared my desk with Youth and Media’s designers, who did an exceptional job and showed me how relevant their work was to prepare an impactful proposal. Rarely have I been as comfortable being challenged and motivated as I was during my time with the Youth and Media team. Sandra always encouraged me to think that many of these ideas were possible in Colombia, and that also motivated me to return to Colombia with this purpose.
BKC’s AI policy practice hosted the recent roundtable discussion about AI and its impacts on youth. What key insights did you gain from that discussion?
The first roundtable we had about the ethical framework was a very interesting discussion. From my perspective, one of the most important points is that experts from all around the world appreciated the effort of including an ethical principle specially dedicated to prioritizing the needs of children and adolescents in the way AI is being developed and implemented in Colombia. I think the Colombian framework will keep this principle and show other countries in the region and the world how important this is.
The discussion also showed me that the impact of AI in the future of education is tremendous. We should be aware of the potential benefits and risks and promote inclusion in AI implementation. AI can accelerate existing disparities and inequalities in this sector and in a country like Colombia, this is something we should definitely avoid.
Finally, there was also a call to consider the need for regulatory measures about AI. I find this interesting because there is no doubt the framework will impact the policy and regulatory proposals we have on this area. At the same time, prioritizing children’s rights and interests will have to be harmonized with existing regulations and international treaties that already developed this same concept.
What advice would you give to someone interested in policy and AI?
I think that someone interested in working in AI policy issues should prepare to be constantly challenged. Not only because concepts are still being developed, but the technology itself is also constantly changing. At the same time, there are high interests at stake that must be considered in every policy analysis. There will be constant pressure for something to be done, and a person working on these areas should be prepared to handle it. I will add that resilience to rapid changes is also a great skill to master if you want to work in this area.
Governments want to show themselves as proactive in this area. When you act as an advisor and say that there is not enough evidence to support a policy or regulatory measure, you will be questioned. Trying to convince decision makers on how important it is to wait, or to develop evidence-based policies, is not an easy task, but it is highly appreciated when this approach provides positive results.
Finally, I think that a multidisciplinary approach is highly desirable. Advising on AI policy requires individuals who can interact with different stakeholders and with technical experts. Presenting yourself as someone with policy expertise but who also understands the technology is helpful. I am a lawyer, and it is interesting to see after a talk or event how some people approach me and consider me a data scientist or computational engineer. I am glad that they see I have this potential and feel confident I understand what they are doing and their motivations.
Are there any resources or additional reading you would recommend to them?
I think that a person interested in these issues should understand the power dynamics involved in this process, the patience that is required to have an impact after a long process of deliberation, and the way policy-making process has been changing in the last decades. Therefore, I recommend them to read these three books:
- AI Superpowers: China, Silicon Valley, and the New World Order by Kai-Fu Lee
- Delayed Response: The Art of Waiting from the Ancient to the Instant World by Jason Farman
- The Cost-Benefit Revolution by Cass R. Sunstein
