Ethics is not the preserve of one continent or culture. Every country, state, government, and agency developing or deploying AI is obliged to do so in line with a sweeping vision that draws upon a geographically, historically, culturally, and socially diverse array of perspectives, particularly those that originate from the crystal of ambiguities that is Latin America and the Caribbean.

Across countries, the impact of AI will not be linear, and within countries, the effect will be uneven. A future that does not belong to you is not freedom. China’s State Council Notice on AI (8th July 2017) highlights China’s ambition to lead the world in state-of-the-art AI technology by 2030.

The six most prestigious taxonomies built in the interest of socially beneficial AI have all emerged from initiatives with a global scope and from within Western democracies. However, Latin American and Caribbean nations must interrogate the differences between AI Preparedness and Network Readiness. They are not logically equivalent.

Network Readiness focuses on factors related to the readiness of a country to harness the benefits of the digital revolution. The Network Readiness Index (NRI 2023) assesses the effect of ICT on society across four dimensions – technology, people, impact, and governance.

However, AI preparedness brings into sharp focus a capabilities and capacity approach to development. The IMF AI Preparedness Index (AIPI) goes beyond network readiness. The AIPI uses a rich set of macro-structural indicators that cover the countries’ digital infrastructure, human capital and labour market policies, innovation and economic integration, and regulatory and ethical frameworks.

The AIPI distinguishes Foundational AI Preparedness (Digital Infrastructure and Human Capital and Labour Market Policies) from Second-generation Preparedness (Innovation and Economic Integration, and Regulation and Ethics).

The AIPI digital infrastructure pillar takes into account the maturity of the private sector’s e-commerce infrastructure, the use of mobile phones for online transactions, the public sector’s online services infrastructure, cost of internet access, the number of wireless subscriptions, estimated internet users per 100 inhabitants, and postal reliability.

The human capital and labour market policies dimension measures the mean years of schooling, gross enrolment ratio, public education expenditure, the digital skillset of graduates in fields such as coding among the active population, the number of STEM graduates, the flexibility of wage determination, pay and productivity, internal labour market mobility, labour market policies including skills matching and retraining, and the percentage of the population covered by social protection schemes.

Innovation and Economic Integration is measured by looking at R&D spending per unit of GDP, frontier technology readiness, AI-related R&D activity, the number of patents on frontier technologies, domestic credit to private sector (%GDP), mean tariff rate, non-tariff barriers, free movement of capital and people, including financial openness and capital controls. Finally, Regulation and Ethics look at the adaptability of the legal frameworks to digital business models, citizen voice and accountability.

The impact of AI is likely to vary across countries at different stages of development, or with different socio-political dispositions. Emerging markets and developing economies may initially face fewer AI-induced disruptions. The AI divide could exacerbate existing economic disparities as advanced economies harness AI for competitive advantage. Emerging markets and developing economies will grapple with assimilating and accommodating AI into their development plans.

The AIPI highlights the urgency for regulatory frameworks to address cyber threats. Where foundational preparedness is weak, investment in digital architectures and human capital development should be prioritised to reap early gains from AI-Assemblages, while laying a foundation for second-generation preparedness.

While the capacity and capability to innovate and strengthen regulatory frameworks for digital businesses are foundational to digital investment in low-income countries, these soft laws can be less effective without strong AI infrastructure and a digitally skilled workforce.

Workers who can harness AI will experience an increase in their productivity and salaries — and those who cannot will fall behind. AI will likely worsen intergenerational immobility and inherited inequality within countries. This requires policymakers to proactively participate in framing ethical AI protocols to prevent AI from stoking social tensions in Latin America and the Caribbean.

Countries may consider establishing comprehensive social safety nets. Employers and bureaucracies can offer training opportunities linked to promotion and salary scales for vulnerable workers. This will make the AI transition inclusive, as livelihoods are sheltered, and increasing inequality is curbed.

Globally, the pattern of adoption and absorption of AI will manifest a slow burn because of substantial costs and investment associated with learning and deployment, and the need for regulation. After the initial slow uptake, the cumulative effect of competition and improvement in complementary capabilities and process innovations will result in sprints.

It is a misjudgment to interpret the early “slow burn” pattern as an indication that the effects of AI will remain narrow. The size of the benefits for the evangelists will cascade in later years at the expense of the antibodies that resist adoption.

If AI meaningfully complements higher-income workers, it may lead to an uneven increase in their labour income. AI will impact income and wealth inequality within countries. As AI evolves, unregulated polarisation within income brackets will accelerate. AI preparedness is therefore essential if we are to harness the potential of AI and lessen its inherent risks.