Latin American Cities in the Age of AI: Navigating the Technological Revolution

By Felipe Suárez Giri

Artificial Intelligence (AI) is transforming cities. It is redefining how governments plan, manage, and police them. It is also reshaping the circumstances under which citizens work, move, and engage in civic life. The long-term impacts of AI in urban environments are still uncertain. However, transformations will likely be profound and enduring as these technologies advance and permeate diverse facets of our lives.

AI has the traits of general-purpose technologies, which are defined by their significant impact on productivity1, widespread application across industries, improvement over time, and generation of complementary innovations2,3. Classic examples are the steam engine, electricity, and information and communications technologies. Through their impact on economic, social, and physical structures, these technologies drastically transform cities and how we live in them.

As Latin Americans, we inhabit one of the world’s most urbanized regions. Considering the remarkable advancements in AI that we are currently witnessing; it is essential to reflect on the benefits and risks these technologies bring to our cities and conceive mechanisms to manage them. This article discusses a few of the multiple potentially disrupting impacts of AI in cities and explores strategies for navigating change.

AI and the Future of Work

Over the last few months, we have seen outstanding progress in generative AI, spearheaded by the release and exponential adoption of Chat GPT. The chatbot achieved the milestone of 100 million monthly active users within the first two months of its release, making it the fastest-growing application in history4. This wave of breakthroughs is creating immense expectations for its capacity to unleash productivity gains. Yet, it is also causing great apprehension over its potential for widespread work automation. According to the last paper released by Open AI, 19% of jobs in the US have at least half of their tasks exposed to AI automation5. Meanwhile, a recent study by Goldman Sachs estimates that 18% of jobs globally could be automated by AI, with Latin American economies facing a higher risk than the global average6.

While there is still significant uncertainty about the likelihood and timing of the impacts of AI in the labor market, these technologies threaten to erode the intrinsically urban base of white-collar jobs7. The studies mentioned above have found that occupations that require tertiary education are particularly vulnerable to automation. Specifically, traditional administrative, legal, sales, finance, and design jobs have been identified as being at the highest risk. Should these predictions materialize, advancements in AI could have regressive effects on income distribution both between and within cities.

On the one hand, capitals and large dynamic cities where talent and corporate headquarters are concentrated, such as São Paulo, Santiago, or Mexico City, will likely attract most of the new high-paying AI-connected positions. By contrast, smaller and less sophisticated metropolitan areas might be unable to attract lucrative jobs. On the other hand, the AI rollout could harm income distribution within cities if a large share of the white-collar workers displaced by automation move into lower-paying roles, hollowing the middle class.

The realization of these scenarios would further reinforce the already high levels of inequality that characterize Latin American cities. From a spatial perspective, the interplay of these factors could exacerbate the stark residential segregation of our cities. Furthermore, it could trigger new inter and extra-regional migration patterns, which are hard to predict now.

The social and economic impacts of work automation will depend on multiple factors. While many of these are beyond government control (i.e., the pace of innovation), policymakers have multiple leavers at their disposal. The early conception and implementation of regulations and safety nets can mitigate the technological transition’s distributive effects. Numerous best practices have been learned from the policy response to the social and economic consequences of the COVID-19 pandemic, including cash transfer schemes, upskilling programs, rent subsidies, and eviction and utility shutoff moratoriums9. All of these could help to temporarily mitigate the adverse effects that automation might have on households’ incomes. However, AI experts warn that structural policies will be required for alleviating the distributive impacts of technological change in the long run. Notably, Sam Altman, CEO of OpenAI, is a vocal proponent of implementing Universal Basic Income (UBI) schemes10. Altman’s argument highlights the importance of UBI in the coming years for safeguarding individuals from the potential decline in salaries resulting from labor market disruptions triggered by AI.

AI and Surveillance

In recent years, the adoption of AI-powered surveillance systems has rapidly increased in Latin America. Argentina, Brazil, Colombia, Ecuador, Mexico, Paraguay, and Peru already have ongoing deployments of facial recognition software11. Considering the high levels of violent crime that characterize many cities in Latam, the allure of these technologies is evident. Nonetheless, the deployment of AI surveillance technologies in the region has been carried out on weak legal grounds, without proper human rights assessments, and has been characterized by its lack of transparency11.

There are two main reasons why the unrestrained rollout of AI surveillance is troubling. First, these technologies have built-in algorithmic and data biases12,13. Commercial facial recognition software, for example, exhibits significant gender and race disparities in its accuracy. Notably, these technologies are particularly inaccurate in the case of dark-skinned individuals and women14. Moreover, algorithms trained on arrest data can exacerbate pre-existing racial biases in law enforcement15. Secondly, without proper safeguards, facial recognition technologies can be a major threat to the rights to privacy, freedom of expression, and peaceful assembly11. A notable illustration of these concerns is the recent judicial investigation which uncovered that Buenos Aires’ surveillance system was unlawfully utilized for obtaining data about journalists, politicians, activists, union leaders, and business people16. Evidently, in the hands of authoritarian-leaning leaders, the misuse of these technologies could put our democracies in peril17.

Whether facial recognition has a place in democratic societies is still hotly contested. However, it’s essential that its use is anchored in human rights in countries where it is already deployed. In this regard, city-level regulations can play a pivotal role in federal systems, by enacting more stringent conditions for the use of AI than those set at the national level11. Some measures which can help minimize the risks to human rights associated with facial recognition surveillance systems include: limiting their use to exceptional cases and requiring judicial authorization, establishing robust oversight and accountability mechanisms, and mandating the disclosure of data about the features and performance of the technologies in use11.

Navigating technological change

AI will continue to permeate and transform cities. It will allow governments to improve citizen engagement, reduce tax fraud, and optimize energy, mobility, and waste management systems, among other benefits18. It will also spark significant productivity gains in the private sector. However, as illustrated above, these benefits come with associated risks. The following guiding principles, derived from the literature and the cases discussed above, can serve as a useful starting point for policymakers in addressing the opportunities and challenges presented by the AI revolution in Latin American cities.

  1. Fairness and equity: AI systems should not discriminate based on personal characteristics19,20.
  2. Transparency and accountability: Governments should be open about the AI systems adopted, the nature of their use, and their performance19,20. A best practice in this dimension, pioneered by Helsinki and Amsterdam, is building algorithm registries where information about the technologies in use is disclosed21,22.
  3. Privacy and security: Adequate safeguards should be adopted to guarantee individual rights and data protection23.
  4. Participation: Community engagement should be considered in cases where AI can impact access to opportunities, rights, or safety19.
  5. Social Protection: Policies to ensure the well-being of households affected by job automation should be devised.
  6. Cooperation: Collaboration between municipalities is critical for adapting AI technologies, which come predominantly from the Global North and China, to address local challenges26.
  7. Proactivity: Cities should adopt a proactive rather than reactive approach to technological change15,19. Integrating AI considerations into strategic planning is a best practice in this dimension18. Other key measures include devising policies aimed at building talent, developing digital infrastructure, fostering R&D, and collecting relevant data 24,25.

The impacts of AI on our economies, societies, and democracies will critically depend on the nature and timing of governments’ responses to technological change. Integrating the principles outlined above into public policies is an initial step towards harnessing the benefits that AI technologies offer, while mitigating their potential negative impacts on equity, human rights, and freedoms. Nevertheless, it is imperative to continue to expand and refine these, as AI continues to advance, and its impacts become more tangible and better understood.


  1. Crafts, N. (2021). Artificial intelligence as a general-purpose technology: an historical perspective. Oxford Review of Economic Policy, 37(3).
  2. Bresnahan, T. F., & Trajtenberg, M. (1995). General purpose technologiesEngines of growth’?. Journal of econometrics, 65(1), 83-108
  3. Lipsey, R. G., Carlaw, K. I., & Bekar, C. T. (2005). Economic transformations: general purpose technologies and long-term economic growth. Oup Oxford.
  4. UBS (2023). Let’s chat about ChatGPT. UBS Wealth Management.
  5. Eloundou, T., Manning, S., Mishkin, P., & Rock, D. (2023). Gpts are gpts: An early look at the labor market impact potential of large language models. arXiv preprint arXiv:2303.10130.
  6. Hatzius, J., Briggs, J., Kodnani, D., & Pierdomenico, G. (2023). The potentially large effects of artificial intelligence on economic growth. Goldman Sachs Economic Research.
  7. Muro, M., Maxim, R., & Whiton, J. (2019). Automation and artificial intelligence: How machines are affecting people and places. Brookings Metropolitan Policy Program.
  8. Busso, M., & Messina, J. (2020). The inequality crisis: Latin America and the Caribbean at the Crossroads. Inter-American Development Bank, 32(10.18235), 0002629.
  9. Bogle, M., & Kumari, S.(2021) Best Practices in Safety Net Programs to Inform an Equitable COVID-19 Recovery, Urban Institute.
  10. Altman, S (2021) Moore’s Law for Everything.
  11. Caeiro, C. (2022). Regulating facial recognition in Latin America: Policy lessons from police surveillance in Buenos Aires and São Paulo [Research Paper]. Royal Institute of International Affairs.
  12. O’neil, C. (2017). Weapons of math destruction: How big data increases inequality and threatens democracy. Crown.
  13. CAF (2021) ExperiencIA. Datos e Inteligencia Artificial en el Sector Público, CAF.
  14. Buolamwini, J., & Gebru, T. (2018). Gender shades: Intersectional accuracy disparities in commercial gender classification. In Conference on fairness, accountability and transparency.
  15. Koseki,S., Jameson, S,. Farnadi,G., Rolnick, D., Régis, C.,& Denis, J. (2022) AI and Cities:Risks, Applications,and Governance. UN-Habitat.
  16. Red en Defensa de los Derechos Digitales (2022) Juez suspende sistema de reconocimiento facial de Buenos Aires por uso ilegal de datos biométricos. R3D.
  17. Keremy, R. (2020). Brazil is sliding into techno-authoritarianism. MIT Technology Review.
  18. Pellegrin, J., Colnot, L., Delponte, L.(2021). Artificial Intelligence and Urban Development. European Parliament, Policy Department for Structural and Cohesion Policies, Brussels.
  19. NYC Mayor’s Office of the Chief Technology Officer (2021) The New York City Artificial Intelligence Strategy, NYC CTO.
  20. OECD AI Policy Observatory (2021) OECD AI Principles, OECD.
  21. City of Helsinki (2023) Artificial intelligence systems of Helsinki.
  22. City of Amsterdam (2023)Algorithmic systems of Amsterdam.
  23. Rodríguez, A.(2021) AI ethics in policy and action: city governance of algorithmic decision systems, CIDOB Briefings.
  24. The Economist (2022) Seizing the opportunity: the future of AI in Latin America,The Economist Group.
  25. Mejía Jaramillo, M. & Torres Paz, J. (2020) Uso responsable de la inteligencia artificial en el sector público. Policy Brief 17. CAF.
  26. KHIPU (2020) Strengthening Artificial Intelligence In Latin America. Outcomes of the first Khipu Latin American meeting in Artificial Intelligence. KHIPU.

Felipe Suárez Giri is a Humphrey-SPURS Fellow at the Massachusetts Institute of Technology (MIT), with expertise in urban policies, infrastructure, and economic development. Prior to MIT, he worked for the Ministry of Housing and Spatial Planning and the Central Bank of Uruguay. He holds a Bachelor’s in Economics, a Master’s in International Development, and another in Urban Studies.