By: Staff Writer
October 23, 2026
The Economic Commission for Latin America and the Caribbean (ECLAC) said in a report on Artificial Intelligence, the Latin American Index of Artificial Intelligence (ILIA 2026) that modern day AI governance is nothing more than ethics washing and not delivering meaningful enough reform and regulations on AI.
The report said the evidence on this is not merely a technical implementation failure that can be corrected with more resources. It reflects a deeper pattern: AI governance has developed as a grammar of principles—a form of “ethics washing,” consisting of high-level statements.
The report also said: “The expansion of artificial intelligence (AI) is altering the conditions under which countries can participate in the digital economy, innovate, and generate value. Connectivity remains a fundamental requirement, but it is no longer sufficient on its own. The development and adoption of AI also require computing infrastructure, data that can be used effectively, specialized talent, scientific and business capabilities, and an institutional framework capable of guiding its deployment, managing risks, and sustaining policies over time.
The Latin American Artificial Intelligence Index (ILIA) analyzes these capabilities across 20 countries in Latin America and the Caribbean, based on three dimensions: Enabling Factors; Research, Development, and Adoption; and Governance. The 2026 edition expands the assessment to 126 sub-indicators—27 of which are new—includes Trinidad and Tobago for the first time, and incorporates a comparative analysis with Germany, Spain, Estonia, Portugal, and Singapore, which helps put some of the strengths and gaps observed in the region into perspective.
The ILIA classifies countries according to their level of maturity across the index’s three dimensions:
• Pioneers: countries with more than 60 points, demonstrating high and relatively well-established performance in infrastructure, talent, research, innovation, and governance.
• Adopters: countries with scores between 35 and 60, which have developed significant capabilities but still face gaps that limit a broader deployment of AI.
• Explorers: countries with fewer than 35 points, whose ecosystems are in the early stages of development, with greater constraints in areas such as infrastructure, research, adoption, or governance.
“As of 2026, Chile, Brazil, and Uruguay are among the “pioneers,” while the “adopters” group—which is larger and more diverse—is led by Colombia, Costa Rica, and Argentina. The inclusion of countries with significant differences in size, population, economic structure, and market scale among the pioneers demonstrates that there is no single path to achieving high levels of AI development. Some countries may leverage the scale of their markets, infrastructure, and productive capacities, while others advance by building on institutional strengths, talent, research, or other enabling conditions.”
The report continued: “The results also show that the nature of these gaps shifts as the technological and organizational demands of AI increase. While connectivity has become widespread, disparities in quality and affordability persist, and the availability of advanced infrastructure—particularly computing resources—remains highly unequal. Although education and research in AI are expanding, the supply of specialized talent and the long-term sustainability of scientific communities remain limited.
“In terms of adoption, progress is faster among individuals than within companies: while the use of AI tools is spreading among the general population, their integration into production processes remains more limited and uneven. Furthermore, within the business sector, adoption rates decline as company size decreases. Regarding governance, strategies and institutional frameworks are becoming more established, while funding and implementation capacity are gaining increasing importance.”
The notion of “responsible AI” has become the primary signifier in global discussions on technological governance; however, its institutionalization is progressing far more rapidly than its actual capacity to protect rights and transform practices. In this chapter, I argue that the criteria for responsible AI cannot be reduced to mere lists of abstract principles—such as fairness, transparency, explainability, and accountability—or to organizational compliance checklists. Such formats presuppose a level of stability in both the technical artifact and the moral context that is, in fact, contradicted by the inherently sociotechnical and evolving nature of AI.
“Global discussions on responsible AI no longer center on whether ethical frameworks are needed, but rather on why these frameworks fail to translate into effective protections—a critique that academia has been substantiating for several years,” the report noted.
It also said: “The Global Index on Responsible AI (GIRAI)—which, in its 2026 edition, evaluates 135 countries based on over 68,000 data points organized into 38 indicators and five dimensions (inclusion and diversity; ethics and sustainability; labour and skills; trust and safety; and the use of AI in public services)—provides the most comprehensive assessment available regarding this gap. Its primary finding: 126 of the 135 countries have at least one AI policy, yet this proliferation of regulations rarely translates into meaningful protection.
“The Global South, for its part, accounts for 203 of the 306 new regulatory frameworks that have emerged since the first edition; however, 78 percent of these lack binding force—compared to 42 percent in the Global North—revealing a structural asymmetry in enforceability between the two regions. It is also telling that, although transparency and explainability are among the most well-developed indicators, only 18 percent require the public disclosure of algorithms used by the state itself.
“Furthermore, the index documents evidence of governments deploying risk-assessment systems that could be considered unacceptable in 35 countries, and it notes that the weakest dimension globally is the use of AI in public services—precisely the area where automated systems affect access to healthcare, housing, education, and migration.”