As cities embark on integrating artificial intelligence (AI) into their infrastructure, a dichotomy emerges: they can either bolster democratic governance or pave the way for surveillance states. Particularly in Latin America, where over 80% of the population resides in urban areas, civic demands for effective solutions to crime and public service inadequacies have surged. Many leaders are turning to AI as a seemingly quick solution that can project a facade of control while sidestepping the challenging, long-term processes of institutional reform.
AI’s Role in Latin America
Latin America’s digital connectivity gap, estimated at $68.5 billion, coupled with a reliance on foreign technological frameworks, often leads cities to adopt ready-made surveillance technologies rather than create systems that center around human rights. This tendency raises alarms as cities become consumers of AI rather than creators of tailored solutions that respect democratic ideals.
The pressing question is not whether Latin American cities should utilize AI but how to implement it in ways that uphold democratic principles and prevent misuse, ultimately serving the public good. Fortunately, the region is already cultivating promising models of AI that reinforce these values.
Navigating the Regulatory Landscape
According to Eduardo Levy Yeyati, there is an urgent need for Latin America to establish its own framework for AI governance. Failure to do so risks turning the region into a "regulatory colony" of tech giants in Silicon Valley or Brussels. In this landscape of diverging U.S. deregulation and strict European regulations, Latin American countries have the chance to carve out a "third way" that nurtures innovation while protecting citizens’ rights.
The urgency for regulation stems from a variety of experiences across the region. From discriminatory facial recognition in Brazil to Argentina’s revolutionary PROMETEA legal automation, Latin America has seen both the advantages and potential authoritarian misuses of AI in governance. Examining these case studies highlights that the region possesses valuable insights into building effective AI frameworks.
AI Innovations Launched in Cities
Latin America faced an AI crossroads even before the advent of large language models like ChatGPT. The technological landscape is multifaceted, sourcing from U.S. and Chinese entities while increasingly fostering local innovations. A notable example is Argentina’s PROMETEA initiative, developed entirely by local attorneys and technologists to automate legal documentation while maintaining human oversight.
In stark contrast, a pilot program launched in Rio de Janeiro featured AI-powered facial recognition technology but faced issues of racial bias and inaccuracy. According to watchdog groups, false positives led to wrongful arrests, particularly targeting Black communities.
In a different scenario, Buenos Aires’ PROMETEA illustrated the advantages of local innovation, streamlining legal work while embedding safeguards against ethical pitfalls. The system dramatically reduced document drafting time while continuously allowing human attorneys to exercise control. Meanwhile, Colombia’s MAIIA platform succeeded in mapping informal settlements, pinpointing infrastructure deficits effectively.
However, conflicting narratives exist. Buenos Aires courts began phasing out PROMETEA, opting for ChatGPT, which raises concerns over the trade-off of transparent, localized systems for proprietary, less accountable technology.
A Call for Human-Centered AI Design
To succeed, AI systems must adopt a human-centered philosophy, complementing rather than supplanting human decision-making. Promising AI projects in the region emphasize transparency and ethical guidelines from the outset. In cases like Argentina’s PROMETEA and MAIIA, human involvement is prioritized throughout the process, illustrating the potential for democratic AI engagement.
Failed initiatives often over-rely on automation and lack transparency, which results in ineffectiveness and even amplifies societal inequities. The PredPol system in Uruguay demonstrates this; despite promising to predict crime locations, it failed to surpass traditional methods and was quickly discontinued. Likewise, questionable projects in Argentina’s Salta province sought misinformed predictive measures regarding teenage pregnancies, breeding concern over privacy and rights infringement.
Promoting Local Solutions
Establishing a framework for democratic AI governance that respects civil liberties is entirely feasible. Immediate safeguards such as transparency requirements, rigorous pilot testing, and the preservation of community-directed AI systems must take precedence. This is not to reject technology but to reinforce democratic values.
Cities must consider additional legislation, similar to those proposed in Chile, to establish registries detailing all AI systems deployed. Focusing on transparency can significantly reduce the risk of authoritarian misuse while promoting authenticity in innovation.
Furthermore, collaborative frameworks like UNESCO’s Ethical Impact Assessment can guide local governance. Regions like Latin America can benefit from strategic partnerships, such as the Inter-American Development Bank’s fAIr LAC initiative, to share best practices and negotiate with technology vendors for fairer pricing and rights protection.
As AI technology evolves, it will ultimately hinge on the policies that define its deployment. The trajectory of this new wave of innovation will determine whether it serves the interests and well-being of citizens or devolves into mechanisms of control. Thus, cities must carefully navigate their AI futures, ensuring they promote democracy and shared governance amid a landscape fraught with challenges.
