Q3 2024 – Artificial Intelligence, data and digitalization

In Argentina, our challenge is to promote digitalization, broaden the data community, foster collaborative data exploration and contribute to the public conversation on these developments, particularly the use and regulation of Artificial Intelligence (AI). This initiative will foster awareness of the potential of data and AI to accelerate progress towards the Sustainable Development Goals agenda. We are currently working on projects that promote digital literacy, particularly the ability to interpret data visualizations. At the same time, we aim to enhance understanding of the opportunities and risks associated with harnessing AI for sustainable development. Our long-term vision is to leverage collective intelligence to discover patterns within the vast ocean of data that surrounds us, generate evidence-based information, and identify future pathways for AI. Ultimately aiming for regulations that ensure a balanced approach, maximizing AI’s potential while mitigating its risks.    To address this challenge, the CoLab has implemented a comprehensive portfolio of actions. It has developed knowledge products exploring the impact of AI on development, highlighting its potential benefits, risks and emerging trends. These resources also analyze the current state of AI-related regulations in Argentina. The CoLab successfully organized seminars, strengthened AI-related capabilities within our Country Office, conducted prospective exercises, and supported a data visualization contest called "Contar con Datos" (which translates to "Tell Stories with Data" and "Count on Data to Make Decisions").  

When this action plan started at the national level, policies and plans aimed at addressing the developments and challenges of Data and AI were, recognizing this critical need, and the request of the resident representative closely monitoring these issues, we are taking proactive steps to fill this gap. Today, the conversation is very productive, with many legislative projects being discussed. The role of the UNDP is more critical than ever, as there are differing perspectives on whether to focus on more regulation or innovation.    Not only are there projects to regulate the use of AI at different levels but also plans to integrate it into legislative and policymaking processes. Since data feeds all AI models, the conversation about AI has highlighted the need to generate, systematize, and make data publicly available. Additionally, discussions are underway to update legislation on protecting personal data.    In today's digital age, we generate an unprecedented amount of data through everyday activities such as using public transportation, making purchases, and interacting on social media. This surge in data production has vast potential to enhance various aspects of life, from improving web navigation and city management to monitoring pandemics and informing public policy. However, the mere availability of data is insufficient. To truly benefit from this wealth of information, it is crucial to know how to process, understand, and communicate it effectively, making ethical use of data and AI.    In Argentina, data literacy remains a significant challenge despite the growing availability of public datasets. Although government institutions, universities, and various organizations have made efforts to publish datasets, many remain underutilized due to limited awareness, lack of training, and low digital literacy across various sectors. This situation is driven by historical gaps in education and the digital divide, which hinder people's ability to engage with and interpret data effectively. The challenge is compounded by a lack of robust initiatives that connect these datasets with actionable insights for decision-making at both local and national levels. As a result, the full potential of data to inform policy, drive innovation, and foster accountability is not being realized. This inefficiency not only limits the effectiveness of public services but also prevents citizens, organizations, and policymakers from leveraging data to address societal issues and improve outcomes. Improving data literacy is critical for enabling people to use data as a tool for informed decision-making, empowerment, and innovation.    Our Lab actively began addressing this challenge last year. Since then, we have made progress on several actions. We have conducted interviews with a diverse group of stakeholders and experts across the country; from universities and foundations to laboratories and companies. These interviews provide valuable insights into the local AI landscape and our country's current position in AI advancement.    Our work strives to mainstream AI within the Country Office. This involves fostering conversations, engaging programme areas, and building staff capacity. To achieve this, we have shared key actions and invited senior management to participate in various activities. Additionally, we organized a training session utilizing AI tools for Country Office staff and consultants.    In addition to the aforementioned actions, we have also published three blogs discussing fundamental AI concepts, Argentinian public opinion on the technology and the debate on regulation. Alongside these blog entries, we published a report and two infographics showcasing the results of our 2023 pilot project integrating Natural Language Processing (NLP) into political speeches analysis.    Both the infographics and the report feature the findings of using a ChatGPT-based NLP model to analyze political speeches. This project aimed to identify shared issues, ideas and concerns within these speeches. As part of the anniversary of the 40 years of democracy in Argentina, leveraging AI tools, the Lab investigated presidential speeches delivered in Argentina since the return of democracy in 1983. The model successfully identified common topics and values among diverse speeches and established correlations between them.     This project served a dual purpose: as a proof of concept for AI’s potential in supporting democratic governance, and as a prototype to explore how language models can foster depolarization. For instance, the learnings could inform the application of these models in future research. Imagine using them to generate fictional text (vignettes) for survey experiments, evaluating voter's ability to distinguish between human-written and computer-generated text. This has significant relevance to studies on the proliferation of fake news during political campaigns.    The Lab’s focus on this topic is particularly timely. Deep political divisions and polarization, evident at both federal and state levels, significantly hinder progress on critical issues. This gridlock fuels public discontent erodes trust in governing institutions and fosters anti-establishment expressions. While overall democratic support remains high compared to other countries, these factors create a breeding ground for the erosion of civic values and governmental legitimacy. Dialogue and consensus building are urgently needed to bridge these divides and address pressing challenges.    To advance our goal of raising awareness and stimulating discussion about the potential and challenges of AI in development, we hosted a free, open seminar on the topic. The event convened experts, practitioners, and policymakers to explore AI’s potential in sectors like health, justice, government and business. Discussions addressed potential challenges and risks.  This well-attended event, drawing over 100 participants, marked a significant achievement in positioning UNDP as a key player in Argentina’s AI conversation. Further extending the reach of our work, Lorena Moscovich, our Head of Experimentation, delivered a TED Talk focused on AI-powered solutions for policy and development, to a capacity crowd of 11000 spectators. This presentation provided a valuable platform to share the insights and learnings garnered throughout 2023.  In this quarter we deepen our reach with the actions and advocacy in the public sphere in our agenda of IA.  In October 2024, we organized a seminar on AI and education at the Ministry of Foreign Affairs, hosted by the G20 Unit of the Ministry, UNDP Argentina, and UNESCO. The event, titled “Artificial Intelligence: New Skills and Horizons for the Future of Education in Argentina”, brought together experts and academics to analyze how AI is transforming various educational contexts, from classrooms to policymakers in Argentina. The discussions focused on exploring the tools and skills individuals need to navigate and benefit from advancements in AI. We are currently working on a report to synthesize the insights and findings from this exchange. While the report is currently internal, it is intended to inform Argentina’s strategy on AI-related topics within the G20 discussions reflecting the country’s commitment to shaping a forward-thinking agenda in this global forum.  Summary of Results of the AI Agenda 2024:  The head of experimentation of the Lab took part in several forums and consultations regarding the development and the use of the AI, such as:  Speaker at the Awareness Session: How Can Artificial Intelligence Transform Legislative Management? National Senate, Buenos Aires, November 12  Speaker at the Annual Seminar CARI 46: Artificial Intelligence from an International Perspective, Argentine Council for International Relations (CARI), Buenos Aires, October.  Participant in the Roundtable: Artificial Intelligence and European Regulation, European Union Delegation in Argentina, Buenos Aires, September.  The podcast “Toma nota. PotenciaIA Humana”: https://creators.spotify.com/pod/show/tomanota/episodes/Todo-Odos--PotencIA-Humana-e2qrcp6.  Speaker at the 7th Annual Educa Innova: Educating for an Uncertain Future: The Challenges of AI, at siglo XXI University, Córdoba, August 23.  Speaker at the First Forum on Generative Artificial Intelligence in Public Management in Ibero-American Cities, Subsecretariat of International Relations, City of Buenos Aires, and Union of Ibero-American Capital Cities (UCCI), Buenos Aires, December 6          Summary of Results of the AI Agenda 2023:    TED Talk  Moscovich, L. (2024, February). Solutions for policy and development with AI [Video]. TED conferences. https://www.youtube.com/watch?v=-_0C5puviuc    Infographics  United Nations Development Programme [UNDP] (2024). The speech of democracy in its 40 years. https://inteligenciaargentina.org/discurso/en/     United Nations Development Programme [UNDP] (2023). Much in common. The topics and speech of democracy in its 40 years. An analysis of presidential speeches through artificial intelligence. https://inteligenciaargentina.org/en/    Report  Moscovich, L.; Vallejo Vera, S.; Alzú, M. S.; Gómez, F.; Zapata, M., and Moreno, M. V. (2023). Much in common. The topics and speech of democracy in its 40 years. An analysis of presidential speeches through artificial intelligence. Buenos Aires: UNDP. Available at: https://www.undp.org/es/argentina/publicaciones/report-ai-argentine-intelligence    Blogs   Zapata, M., Moscovich, L., Moreno, M.V. & López, M. E. (2024, June 15). Artificial Intelligence, Beyond Algorithms: An Evolving Regulatory Dilemma. Available at https://www.undp.org/es/argentina/blog/artificial-intelligence-beyond-algorithms-evolving-regulatory-dilemma     Moscovich, L. and De Rosa, A. (2023, October 9). Episode II. What do Argentines talk about when we talk about AI? (an example). UNDP Argentina. Available at https://www.undp.org/es/argentina/blog/episode-ii-what-do-argentines-talk-about-when-we-talk-about-ai-example    Moscovich, L. (2023, September 1). AI: Argentine Intelligence. A Conversation about Artificial Intelligence: Episode I: Anyone Can Cook, and Use AI! UNDP Argentina. Available at https://www.undp.org/es/argentina/blog/ai-argentine-intelligence-conversation-about-artificial-intelligence    Seminar: Seminar hosted by UNDP Argentina together with the MAPP of UdeSA in November 2023, titled: "Artificial Intelligence for Development" https://www.youtube.com/watch?v=iY2RqcFMqG8&t=15s    Poster at the Scientific Symposium: Moscovich, L.; Vallejo, S.; Zapata, M.; Alzú, M. S.; Gomez, F., and Moreno, M. V. (2023, October 4). Much in Common: NLP Applications for Depolarization Strategies. The Case of "The Discourse of Argentine Democracy in its 40 Years." Poster presented at the Scientific Symposium on AI and Applications, Buenos Aires, Argentina.    AI Training for the Country Office: In November 2023, a training session was facilitated on different Artificial Intelligence tools for all UNDP staff. The training was led by Dr. Tomás Balmaceda, a technology specialist. 

The value proposition that our Accelerator Lab aligns with the values of the Digital 4 Development Hub in Latin America and the Caribbean. Specifically, we prioritize enhancing quality of life through technology and fostering citizen participation. By cultivating a people-centered digital culture, we aim to make public data more accessible and provide digital platforms for citizens to engage, ensuring that technology serves as a powerful tool for economic growth, creativity, and community building.   Also, the CoLab, together with UNDP, are uniquely positioned to lead the public conversation on Data and AI use and regulation in the country, while promoting awareness of its benefits and risks for development and democracy. The UNDP’s strong reputation and history of convening stakeholders during critical moments make it a natural facilitator. For instance, during the 2001 economic and social crisis, UNDP played a central role in fostering dialogue and consensus-building around shared priorities. Today, while the context is different, UNDP can leverage its experience to identify common concerns and values that can serve as a foundation for a new national consensus on Data and AI, one that reflects the core priorities of the Argentine citizens. Thus, UNDP is a neutral mediator beyond the political divide.   The change in national administration in 2023 presented an opportunity for the Lab to serve as a valuable partner. Given the limited experience with AI initiatives within the new government, the Lab’s knowledge and capabilities can help bridge potential gaps.   As a leader in innovation, we have been exploring various AI applications in areas like elections, democracy, and development. We actively share our findings with the Country Office and aid in enhancing its capacity to leverage the advantages of AI. Furthermore, the Lab has been successful in executing impactful actions, including the pilot project on NLP, the seminar hosted at the end of last year and the TED Talk facilitated by our Head of Experimentation, Lorena Moscovich.   This year 2024, significant progress was made in the field of AI, data and digitalization. A seminar was successfully held in collaboration with the Ministry of Foreign Affairs, a summary of our AI research findings was presented at Universidad Siglo XXI in Córdoba The 'Contar con Datos' contest was also organized in partnership with the Data Science Lab of the University of San Andrés, and the Undersecretariat of Science of the Nation. This data visualization contest expanded the data community and fostered collaborative data exploration. 

Join us to discuss #Data and #AI use and regulation in Argentina. Can it boost sustainable development? Are there risks? We use #CollectiveIntelligence and #DataVisualization to turn daily data into insights, driving evidence-based decisions and fostering public conversation. 

Learning questions that arise from our challenge:  - What are the elements, processes, institutions, and models involved in the regulation of AI worldwide? Is it feasible to replicate these in Argentina? Has the country its own laws and experiences on AI?  - What are the characteristics of AI usage and access in Argentina? Are these compatible with ethical, sustainable, and responsible standards?   - Who are the main players within the national AI ecosystem?  - Can a data visualization contest serve as a tool for collective intelligence to explore publicly available data in Argentina and identify positive deviants?   - What are the community’s areas of interest regarding data?   - Can such a contest help foster data appropriation and enhance data visualization literacy?      To meet this challenge and its related learning questions, the Lab has drafted a portfolio of actions.     This year, we published a blog post that analyzed the current state of AI-related laws and regulations in Argentina and successfully organized a seminar in collaboration with the Ministry of Foreign Affairs.    Regarding data and digitalization, we presented a summary of our AI research findings at Universidad Siglo XXI in Córdoba Additionally, we organized the 'Contar con Datos' contest in partnership with the Data Science Lab of the University of San Andrés, in collaboration with the Undersecretariat of Science of the Nation.     The Lab will further undertake a prospective analysis to identify signals of change that allow us to explore emerging trends and development regarding the AI for Development Agenda. This analysis will encompass the exploration systematic examination of relevant signals. 

The action plan on AI, data and digitalization employs a comprehensive methodology, including stakeholders mapping, interviews with key informants, blogs publishing, application of AI models, collective data generation, data visualization, data contest and the convening of seminars. The long-term ambition of this multi-pronged approach is to leverage collective intelligence to foster a deeper understanding of how to use and regulate data and AI, ensuring a balanced approach that maximizes its potential benefits while mitigating associated risks.    Furthermore, the Lab’s exploration of AI applications within the context of political discourse extends beyond a mere proof-of-concept. It serves as a prototype for testing the potential of AI to contribute to strengthen democratic governance. As previously mentioned, a Natural Language Processing (NLP) model using Chat GPT was developed to analyze political speeches. This model successfully identified recurring themes and shared values among diverse speeches, establishing correlative relationships. Additionally, the model processed real political speeches to generate artificial ones. The pilot project serves as a compelling demonstration of AI’s potential to identify unifying elements within a nation’s political history, showcasing a responsible application of this technology that aligns with democratic principles.    The pilot project holds significant value for further research on the dissemination of misinformation and “fake news” within democratic environments. The extracted insights could facilitate the application of these models in generating fictitious text for survey-based experiments (vignettes) and assessing a voter's capacity to discern between human-generated and computer-generated text.    Furthermore, the Lab leverages AI-powered analytics to examine social media data, aiming to understand public sentiment and opinions on AI within Argentina. A collaborative effort with Citibeats, a social listening company, resulted in the analysis of nearly 22 million tweets from public accounts georeferenced in Argentina during April,2023. This data provides valuable insights into public perception.    The Lab’s approach to fostering collective intelligence involves in-depth interviews with a diverse pool of experts on the cutting-edge applications and regulations of AI for both democratic governance and development. These interviews facilitate the gathering of insights and case studies regarding the potential benefits and risks associated with AI. The knowledge acquired serves as a foundation for establishing categories and variables for analyzing AI aspects relevant to the Lab’s core challenge.    Finally, the Lab hosted a two seminars on AI that fostered open exchange with a broad audience and took part in several forums of discussions 

Discussions regarding the ethical consideration surrounding the use of data, as well as the development and deployment of AI are ongoing. However, these discussions often lack focus on critical aspects, such as:  Promoting digital literacy and the ability to interpret data.   Identifying key players and stakeholders within the national AI ecosystem, including their respective roles and agendas.  Examining potential regulatory models that can support government authorities in fostering ethical, responsible, and secure data use and AI development  Analyzing the pressing challenges that AI presents for Argentina, particularly in the context of its development efforts 

As we move forward within our challenge, the Lab will continue to work in fostering an open and inclusive public debate on AI's applications and regulations. We aim to raise awareness and understanding of AI's potential to contribute to development goals, while also evaluating existing regulatory frameworks to ensure they balance innovation, safety and ethical considerations.      Collaboration with experts and partners has proven instrumental in mapping the national AI ecosystem. We hope that the ongoing exploratory conversations officials in Open Data and Public Sector Innovation allow us to foster partnerships to scale up our efforts. With our contributions, we aim to nurture the AI national ecosystem. By highlighting key experts, projects and organizations, and by connecting stakeholders across the ecosystem, the Lab’s work is laying the foundations for a robust national AI landscape.       Last year, active engagement with our colleagues within the Country Office was a hallmark of our work. We encourage open dialogue and feedback, fostering participation in the challenge. We also organized training sessions to enhance the Country Office’s capabilities in approaching AI. Building on this success, we will continue to share progress and results with our colleagues, and plan to organize another training session to explore how AI can enrich the Country Office’s work.