As artificial intelligence (AI) expands globally, its effects on democracy, human rights, and social justice have become critical topics. The challenges and opportunities brought by technology’s evolution are not the same for societies across the globe. Countries in the Global South face particular issues, from the risks of increasing historical inequalities to the threats of digital surveillance by governments.
On 24 September, 2024, more than 20 experts from Latin America, Europe, and the U.S. gathered at the DemocracIA Summit in Buenos Aires, Argentina, and spoke to an audience of more than 200 people from public and private sectors, civil society, media, and academia. The event was organised by Luminate, Civic Compass, and the International Fund for Public Interest Media (IFPIM) to convene high-level conversations on AI’s governance, responsible innovation, and the impacts of this technology on elections – all with a focus on Latin America's context.
Urgent need for AI regulation
The shared understanding is that the window to regulate AI and ensure it’s in service of the public interest is closing. For Latin America and the Global South, failure to act quickly could worsen existing social challenges, rather than leveraging AI to address them. Latin America, as the most unequal region globally, has the chance to lead a human rights-centred technological revolution – but time is running out.
Bridging the Global North and South in AI regulation
The first panel explored the divide between the Global North and South in AI regulation. Experts warned of the risks of trying to replicate regulatory frameworks designed in Europe and in the U.S. without taking local contexts into consideration.
Horrara Moreira, Project Management Officer at the Center for Exponential Changes at the Beja Institute in Brazil, highlighted the importance of ensuring that AI regulations consider the effects on society's most vulnerable groups: "The concentration of corporate power affects populations differently. A key concern with the AI regulation proposal currently advancing in the Brazilian Congress is its potential impact on poor and Black communities."
Maroussia Lévesque, a professor at Harvard Law School, pointed out that big companies prioritise profit over safety to dominate the market, reinforcing the need for regulations that can ensure those tools are designed and deployed in service of public interests.
“
“Big companies are prioritising speed over safety because the winners in this race for innovation take the largest market share,” said Lévesque.
Fanny Hidvegi from AI Collaborative observed that while there are numerous discussions about the negative aspects of AI we want to avoid, there are few positive visions for how AI should be used. She emphasised the need to identify specific rights that should be central to shaping an affirmative vision for AI as a tool.
Gabriela Hadid, Luminate’s Director for Latin America, moderated the panel, highlighting the importance of including voices from the Global South in international AI governance discussions. Vanina Martínez from the Barcelona Institute for Artificial Intelligence Research warned that excluding these voices leaves a large part of the world behind.
“
“We risk prioritising the interests of politically and economically powerful countries over humanity,” Martínez cautioned.
She called for a global dialogue on AI governance under the United Nations framework, signaling a critical step towards more inclusive agreements.
Where innovation meets responsibility
A second panel, moderated by Javier Pallero from Civic Compass, zeroed in on AI’s ethical implications. Claudia López, from Chile’s National Center for Artificial Intelligence, emphasised the need for transparency about AI’s limitations and urged Global South countries to transition from being primarily consumers of AI technology to developing and controlling their own AI infrastructure.
“We must rethink how we distribute the benefits and costs of AI,” López asserted.
Luciana Benotti, from the National University of Córdoba, explored the concept of "emerging biases" in AI, where systems trained in one context produce harmful outcomes when applied elsewhere. She stressed the need for participatory design and error analysis, specific to different social groups to mitigate these biases.
“Emerging biases in content moderation systems can result in hate speech,” explained Benotti.
Paola Ricaurte, from Harvard’s Berkman Klein Center, warned of the high costs Latin America pays for AI technologies developed elsewhere. She emphasised that while the region provides data, labor, and natural resources, the benefits remain concentrated in the Global North.
“
“If we don’t radically change the conditions of this game, we will see an increase in inequality both globally and in Latin America,” Ricaurte cautioned.
AI and election integrity: A gendered perspective
The final panel, moderated by Julia Pomares from the Global Initiative for Digital Governance, explored AI’s influence on elections. Menno Cox, from the European Commission, discussed how the EU’s Digital Services Act (DSA) was important to curb disinformation during this year’s campaigns: "Thanks to this new regulation, companies within the European Union are now obligated to respond to information access requests and are held accountable to the Commission."
Fernanda Martins, from InternetLab in Brazil, addressed “gendered disinformation,” highlighting how AI-driven disinformation campaigns disproportionately target female candidates. She pointed to sexualised deepfake videos as a particularly harmful form of violence against women in politics.
“We must recognize that AI-fueled electoral disinformation affects female candidates more severely,” Martins noted.
The panel also examined how journalism can counteract AI-driven disinformation. As AI-generated content becomes more common, accurate and trustworthy journalism will be crucial in detecting and combating false narratives.
The road ahead for Latin America
The discussions at the DemocracIA Summit served as a wake-up call for Latin America to shape its AI future. Without inclusive and rights-based regulation, AI risks deepening the region’s entrenched inequalities. However, if used responsibly, it holds the potential to promote social equity and strengthen democratic governance.
The clock is ticking, and now is the moment to ensure AI serves democracy before it’s too late.