Operational regulation: Strengthening AI governance in Latin America

Recent elections in Latin America have highlighted the dangerous influence of algorithms and artificial intelligence, especially generative artificial intelligence, at polling stations. During Colombia’s 2023 regional elections, campaigns were tainted by disinformation spread through social media, with the emergence of artificial intelligence-generated media content – or deepfakes – aimed at undermining candidates’ political campaigns. This has raised concerns among fact-checking groups like ColombiaCheck because the widespread and convincingly realistic recordings enabled by the imitation generative artificial intelligence have proven too difficult and time-consuming to disprove before they do any damage. A similar fraud occurred in Argentina’s 2023 elections, where candidate Patricia Bullrich was the target of AI-generated audio recordings defaming her economy minister. During Mexico’s recent elections, candidate Claudia Sheinbaum’s deepfake audio mimicked her criticism of President Andrés Manuel López and admission of poll fraud.

Political actors are undoubtedly using AI, for example by exploiting its ability to “microtarget” voters. This raises serious privacy and ethical issues. Disinformation, targeted advertising, and the resulting “filter bubbles” are on the rise, contributing to political polarization and voter manipulation. In parallel, the use of AI-powered bots and cyborgs designed to imitate humans and interact with targeted media users to promote the spread of disinformation is widespread. Deepfakes and bots plagued Brazil’s 2022 presidential election. In Costa Rica’s 2022 presidential election, more than 11 percent of social media posts came from fake accounts, rising to 19-21 percent in the weeks leading up to the vote.

This in itself is nothing extraordinary. What sets Latin America apart, however, is the unique challenges the region already faces. Besieged by political instability, polarization, and fragile institutions, Latin American democracies are on fragile ground. Combined with systemic corruption, unstable governments, and poor digital literacy, this state of affairs has hampered the development of comprehensive regulatory frameworks and delayed the emergence of regional initiatives to combat the misuse of AI. In reality, AI-based technology operates in Latin America without stringent regulations, allowing political actors to disseminate vast amounts of disinformation to the electorate with impunity.

As the past year has shown, AI has ushered in a new era of political campaigning, threatening the region’s fragile democratic stability. Cooperation with the European Union, which is already a step ahead in regulations and good practices, is crucial to rethink and protect electoral processes in Latin America and prevent the degradation of democratic spaces in the digital sphere.

Indeed, the EU is positioning itself as a global leader in AI regulation, with an emphasis on ethical standards, data privacy and human rights. For the EU to maintain its leadership and ensure global digital stability, it must extend these regulatory rules beyond its borders. By helping Latin America develop strong AI regulations, the EU can promote a coherent global approach to AI governance, ensuring that ethical and security standards are adopted and maintained around the world.

In addition, the EU has a strong interest in supporting the health of democracy in the region. Democratic instability in Latin America has already caused increased migration, trade disruptions and security threats that directly affect Europe, while also straining relations with important EU partner countries. Economically, the EU is one of the largest investors in Latin America, with EU foreign direct investment stocks in the region expected to reach €680 billion from 2022 (including the Caribbean). For example, Spanish telecommunications giant Telefónica has significant operations in Brazil and other Latin American countries, where a stable regulatory environment is key to ease of doing business. Once finalised, the EU-Mercosur trade agreement will create one of the world’s largest free trade areas, underlining the importance of uniform regulatory standards for smooth economic relations.

To reduce the negative impact of generative AI, countries and multilateral forums such as the United Nations, G7 and OCDE, as well as the EU, have adopted ethical guidelines and regulatory frameworks at the national level. These multi-stakeholder efforts aim to guide ethical AI development while safeguarding user rights, with a focus on data protection and reducing data bias in algorithms. At the same time, efforts are being made to welcome the invention of counter-technologies such as blockchain that can mitigate and detect fraud and other threats posed by generative AI. Latin America can learn from these policies.

However, there are still many steps that need to be taken for Latin America to achieve a common approach to effectively governing AI. One of the main challenges is digital literacy, which is exacerbated by stark rural and gender gaps in the region. This, combined with political turmoil, has hampered the implementation and uptake of regional initiatives aimed at responding to the significant risks that AI poses to electoral processes in Latin America. It is therefore important for the governments of the region to take advantage of the renewed momentum of the strategic partnership with the EU, as evidenced by last year’s Digital Alliance Summit between the EU and Latin America and the Caribbean, which became a breeding ground for bilateral initiatives and promises of more structured cooperation. The UN notes that Latin America must start working together to promote and implement ethical frameworks and best practices on artificial intelligence in relation to electoral platforms and processes for the sake of democratic accountability. EU-Latin America cooperation on AI governance can therefore use the EU AI Act as inspiration and support regional progress through organizations such as the Development Bank of Latin America and the Caribbean and the Inter-American Development Bank. This multilateral collaboration can help establish consistent frameworks and standards across the region. It will also need to be tailored to the specific legal and regulatory context of individual Latin American countries, requiring each nation to adapt and incorporate these guidelines into its local legislation to effectively address unique challenges and priorities. In addition, the EU should use its strategic partnership with the region to strengthen capacity building, create joint research initiatives and provide technical assistance. Europe has compelling reasons to be concerned about AI regulation in Latin America. By actively helping Latin America create better AI regulations, the EU can protect its strategic interests, enhance global security and promote ethical standards around the world. Building on its relationship with the region, the EU has many opportunities to work with Latin American governments to ensure that AI is developed and used in a way that benefits both regions and contributes to a stable and secure global environment. With Venezuela, Brazil, Chile and Uruguay still facing elections this year, there is no better time than now.