Elections, climate change, and Artificial Intelligence were the main thematic focuses of Media and Democracy in 2024

By: FGV Comunicação

Contemporary transformations in the political, environmental, and technological spheres have created interconnected challenges that deeply impact society. The current scenario requires comprehensive reflection and concrete action to address issues such as the fragility of democratic systems, the global climate crisis, and the increasing use of Artificial Intelligence-based technologies. These three themes – elections, climate change, and Artificial Intelligence – were at the core of Mídia e Democracia's activities throughout 2024, including the production of 41 studies, 33 misinformation checks, and 8 cycles of the Mídia e Democracia Council. This annual report outlines the key highlights of the project based on these central themes.

The growing misinformation and attacks on the integrity of electoral processes, fueled by the proliferation of misleading content, undermine public trust and compromise democratic representation. On the other hand, climate change continues to intensify socioeconomic inequalities, disproportionately affecting vulnerable populations and demanding urgent responses from governments, businesses, and civil society. Simultaneously, advancements in Artificial Intelligence, while offering potential for significant innovations, raise ethical concerns regarding privacy, discrimination, and the concentration of power in the hands of a few.

This report seeks to explore the intersections and singularities of these three themes, highlighting how the crises and opportunities in each field connect in a global landscape marked by uncertainty. Furthermore, it investigates how societies have responded to these challenges, proposing a critical perspective on the solutions offered and the risks posed by the lack of regulation and coordination between global and local actors. By integrating perspectives from various areas, we aim to contribute to a deeper, interdisciplinary understanding of the digital debate that directly impacts democracy, shedding light on potential paths for promoting social justice, sustainability, and responsible innovation.

We present this policy paper, beginning with the section “Analysis Lines: The Panorama of the Period,” which includes the findings from studies on the three aforementioned themes. These results, along with the analytical dive undertaken to produce these studies, led to the identification of possible lines of action, outlined in the section “ Action Lines: Thematic Directions.” Finally, in the “Retrospective” section, we summarize the main activities carried out by the project throughout 2024.

1. Analysis Lines: The 2024 Overview

In 2024, the project organized its studies around three central thematic axes: elections, climate change, and Artificial Intelligence. Although these themes are diverse, they share a common point: their relevance to the consolidation of democratic practices in an increasingly digital-mediated world. The selection of these axes reflects the need to understand how public debates on these topics develop on digital platforms and how this directly impacts the exercise of citizenship, governance, and collective rights.

1.1 Elections

As observed in the studies on elections conducted throughout the year, the debate on this topic illustrates how social media has become a decisive arena for political communication, where political actors and ordinary citizens construct narratives and vie for visibility around agendas, ideologies, and different worldviews. It is important to highlight that, within these dynamics, several communication challenges are brought to the forefront, such as disinformation, hate speech, and gender-based political violence, directly impacting the quality of public debate and the health of democracies. In the realm of democratic elections, the Mídia e Democracia project focused especially on Brazilian municipal elections but also covered aspects of the U.S. presidential elections and, in a more specific scope, elections in other countries affected by the use of Artificial Intelligence, as will be detailed in a later section.

1.1.1 Approaches

The 2024 municipal elections were analyzed through different lenses within the Mídia e Democracia project, covering key issues essential to understanding the contemporary Brazilian political landscape. Among the focal points of analysis were gender-based political violence against female candidates, which exposes structural inequalities in women's access to and participation in politics; the digital reach of candidates, highlighting strategies and tools used to amplify their visibility on social media; toxicity in candidates' speeches during televised debates, revealing patterns of aggression and targeted attacks; and the boosting of posts on digital platforms, showcasing the growing importance of financing for the success of online campaigns. This multifaceted approach aimed to map the power and communication dynamics in the electoral process, with a special focus on the changes brought about by social media.

The study "Municipal Elections on the Networks: Overview of the First Month of Electoral Dispute for the Mayoral Elections in Brazilian Capitals", for example, presented a general view of the municipal race in the capitals during the first month of the campaign, based on a broad mapping of strategies, approaches, and the candidates' actual reach on digital platforms. The research found that Instagram was the main engagement platform, accounting for 91.4% of the interactions observed, with a focus on videos in the Reels format (88.2% of the most impactful posts). Additionally, it was observed that regional and gender disparities marked the race, as candidates from São Paulo concentrated 81.6% of the total engagement, while female candidates performed less, reflecting structural inequalities. In terms of themes, religiosity and family dominated posts (34%), followed by sports, political violence, and anti-system rhetoric.

One standout element was the issue of candidates who are digital influencers and/or coaches linked to entrepreneurship discourses, such as the case of Pablo Marçal, a candidate for the São Paulo Mayoralty from PRTB, who alone accounted for 80% of interactions considering all the candidates' networks across all the capitals. A similar phenomenon occurred in Fortaleza, as demonstrated by another study within the project: candidate André Fernandes from PL, with informal and humorous discourse, established a direct dialogue with the language of young digital influence and easily dominated the major networks – although, like Marçal, he did not win the election.

Exploring the realm of gender-based political violence, another study pointed out that female candidates were more likely to be targets of aggressive and misogynistic comments (57.7%), while male candidates received more disagreements (32.7%). While men faced irony and derogatory nicknames (16.54% of the comments), women faced more severe attacks, including the disqualification of their abilities (83.46%). Right-wing female candidates, such as Cristina Graeml (PMB-PR) and Emília Corrêa (PL-SE), received higher volumes of support, with 96.8% and 70.4% of favorable comments, respectively. On the other hand, left-wing female candidates, such as Duda Salabert (PDT-MG) and Tabata Amaral (PSB-SP), were the targets of 86% of the recorded attacks. Duda Salabert was particularly targeted with transphobic attacks. The results highlight gender-based political violence in the online environment, with more intense attacks against women from progressive political spectrums and marginalized groups, reflecting structural barriers to female political participation.

Other approaches addressed analyses of the toxicity level of candidates' statements during televised debates, specifics of post-boosting, and the Brazilian debate on the U.S. elections. In the toxicity report, candidates' statements from the city of São Paulo were assessed and classified based on the toxicity level using an AI tool. It was found, for instance, that incidents of physical and verbal violence were preceded by toxic rhetoric, suggesting a link between discourse and aggression. The research on post-boosting revealed that, with Google's ban, Meta (Facebook and Instagram) became the only available platform for paid electoral advertising. In 2024, spending on boosting during the first round surpassed the total in 2020, reaching R$ 194 million. Regarding the U.S. elections, another study showed that comments on the then-candidates Kamala Harris and Donald Trump were polarized, with attacks on the press and conspiracy theories, highlighting racism and misogyny toward the Democratic candidate.

1.1.2 DRI Risk Analysis

The report "Electoral Integrity in the Digital Environment: Risks and Recommendations for the 2024 Brazilian Municipal Elections" was one of the highlights of the project, in its effort to understand the challenges of the Brazilian political landscape in the context of digital campaigns. The document provides a comprehensive analysis of the vulnerability of the 2024 municipal elections to online threats, focusing on three main areas: hate speech, disinformation, and political propaganda. To address these risks, the report presents recommendations targeted at different actors in the electoral process. The report was produced by Democracy Reporting International (DRI), with contributions from FGV Comunicação Rio, FGV Direito Rio, and Lupa, in addition to input from experts from various sectors who participated in roundtable discussions of the Digital Democracy Council.

For electoral authorities, the report recommended strengthening oversight and promoting specific training for responsible bodies, as well as initiatives that encourage transparency in the electoral process and civic education. For digital platforms, the recommendations included the use of advanced tools to identify and mitigate risks, as well as adjusting algorithms to reduce the visibility of harmful content. Finally, civil society organizations were encouraged to lead efforts for stricter regulations in the digital environment and to develop ongoing research to understand the impacts of online threats.

In addition to the report, DRI launched an interactive panel that monitored discourse on Instagram about the 2024 municipal elections in four Brazilian capitals: Rio de Janeiro, São Paulo, Porto Alegre, and Recife. This research investigated the dominant narratives surrounding the main mayoral candidates in each city, providing a detailed panorama of the debates that shaped the digital landscape. With these initiatives, the project aimed to contribute to the creation of a safer and more democratic electoral environment, reinforcing the need for joint actions between public institutions, civil society, and technology companies.

1.2 Climate change

The digital debate on climate change was central to Mídia e Democracia in 2024. The studies produced by the project found a high circulation of denialist and disinformative discourses that could put the population at real risk, especially in the context of extreme events such as the floods in Rio Grande do Sul, which occurred in May of this year. The increased circulation of digital content related to climate, with posts from politicians, activists, scientists, and citizens, highlights the need to understand how these discourses are formed, disseminated, and consumed. Investigating the digital debate on climate change allows for the identification of key actors involved, the strategies used to spread false or manipulated information, and the consequences of this for civic engagement and public policies.

1.2.1 Approaches

In the reports on this theme, we sought to understand, among other approaches, the impact of climate tragedies on the discourse of politicians during municipal elections, the tensions between public and private institutions in the context of extreme events, and the debate on environmental racism. The studies highlight various dynamics, such as the polarization of discussions, the role of influencers and celebrities in digital mobilization during extreme events, and the circulation of disinformation and climate denialism in contexts of high political polarization.

From a transversal perspective connecting the theme of elections with climate change, one of the studies analyzed how the candidates for the mayoralty of Porto Alegre addressed climate issues in their electoral campaigns. During the flooding crisis, around 41% of the candidates' posts directly addressed the issue. However, the debate was more reactive than proactive, focusing on the immediate impacts of the floods rather than on long-term climate policies. The decrease in the volume of posts on the subject in the following weeks suggests that the candidates may have mobilized the climate issue opportunistically, adjusting to public concern without prioritizing a long-term sustainable agenda.

Another angle explored in the project was environmental racism, which was investigated in the context of two major national climate tragedies: the floods in Rio de Janeiro and Rio Grande do Sul, which occurred in January and May 2024, respectively. In the case of Rio de Janeiro, the research identified that while the concept of environmental racism sparked debates and defense from public figures, it also provoked racist and misogynistic attacks, with irony and stigmatization of vulnerable populations, as well as attacks on the Minister for Racial Equality, Anielle Franco, who mobilized the term at the time. In Rio Grande do Sul, the study focused on public reactions to the announcement of policies from the ministry aimed at vulnerable groups, identifying that the predominant narrative strategy was invisibilization, questioning policies focused on the populations most affected.

It is possible to argue that both studies on environmental racism highlighted the manipulation of the debate on social media, with an emphasis on disqualifying reparative actions and minimizing issues of socio-environmental inequality. The presence of stereotypes, online racism, and political attacks reveals the resistance to recognizing the impact of climate change on the most vulnerable populations, particularly black and peripheral communities.

Another perspective proposed centered around the digital narratives about the floods in southern Brazil in the early days of the tragedy, from May 1 to 15, 2024. The analysis revealed significant polarization in the discussions, with a focus on six groups: right-wing, left-wing, influencers, donation organizers, entertainment pages, and ordinary profiles. On Instagram, the most impactful posts were made by celebrities and influencers such as DJ Alok and Whindersson Nunes, who used their platforms to mobilize solidarity. On Facebook, news pages, party-affiliated vehicles, and local content predominated, with a more practical and informative approach to the crisis.

1.3 Artificial intelligence

Throughout 2024, Artificial Intelligence was also a central focus in the Mídia e Democracia project, through studies that illustrate how this technology has become a key element in contemporary communication dynamics. The goal was to highlight different aspects of this topic, demonstrating how AI not only enhances the capacity for content production and dissemination but also intensifies challenges related to misinformation, the creation of deepfakes, the reinforcement of discriminatory biases, among others.

1.3.1 Approaches

Artificial Intelligence (AI) was analyzed in various ways, with a focus on its implications for communication dynamics, electoral processes, and informational integrity. In the broader context of communication on digital platforms, the study Imaginaries about Artificial Intelligence in the Brazilian Digital Debate revealed how this technology has become a central topic on social media, reflecting a range of perceptions from utilitarian optimism to ethical and social concerns. From January to May 2024, over 365,000 posts were collected from Facebook, X, and Instagram. The predominance of positive and neutral imaginaries stands out: 41.02% of the posts evaluated presented an optimistic tone, while 42.36% adopted a neutral or balanced approach. Optimistic content, driven by influencers and companies, reinforces the perception of AI as a solution to everyday challenges in areas like healthcare, marketing, and entertainment. On the other hand, only 7.77% of the posts presented a pessimistic perspective, criticizing the risks of automation and the impact of AI on labor relations and ethics.

Given the relevance of this technology in political contexts, some studies in the Mídia e Democracia project took a more specific approach, analyzing the relationship between AI and elections. Since, in 2024, more than a quarter of the world's population went to the polls to elect their representatives, these studies mapped not only the Brazilian context but also the situations in India, South Africa, Mexico, and the European Parliament. 

Thus, we mapped extensive use of AI for creating deepfakes, either to manipulate images and produce content aimed at spreading disinformation or to attack reputations and influence voters. Cases like the deepfakes of deceased Indian political figures and fraudulent videos in South Africa, involving international personalities such as Donald Trump and Eminem, demonstrate AI's ability to shape local narratives with global impacts.

In this sense, while Brazil and the European Union advanced specific regulations to mitigate these risks, other contexts, like India’s, demonstrated the need for more comprehensive and preventive measures to address emerging challenges. In Brazil, despite the anticipated widespread use of AI in municipal elections not materializing, cases of humorous deepfakes with a "rough" aesthetic and deepnudes targeting female candidates highlighted how AI can amplify preexisting sociocultural dynamics, such as political violence against women and the "meme culture." Beyond the challenges, the studies also highlighted AI’s positive potential for promoting citizenship and democratic integrity. For instance, the use of chatbots by electoral campaigns and institutional bodies demonstrated how AI can facilitate access to information and increase political engagement. 

In addition to the debate on AI usage, the project employed innovative methodologies to capture what the main generative AIs – Copilot (Microsoft), ChatGPT 4.0 and 4.0-mini (OpenAI), Claude 3.5 Sonnet (Anthropic), and Gemini (Google) – had to say about some electoral processes in Brazil and worldwide. This research effort revealed limitations in the accuracy of the information provided, with factual errors and inconsistencies noted – reinforcing the need for regulations and the development of tools that ensure AI's reliability in disseminating information in critical contexts like elections and digital communication.

2. Action Lines: Thematic Directions

The studies conducted throughout 2024 highlight that recurring challenges in the national political spectrum are reflected both in elections and in agendas related to climate change and Artificial Intelligence. In the electoral field, issues emerge involving the behavior of candidates, voter participation, and the growing impact of digital platforms in mediating political dialogue. These factors demand not only the overcoming of structural obstacles but also a consensus on the regulation of big techs, which amplify problems such as disinformation and hate speech.

In parallel, the integration of climate issues into political agendas reveals a scenario of urgency. The lack of political commitment to mitigating and adapting to climate change has direct consequences on the quality of life of the population and the country's sustainability. Meanwhile, advancements in Artificial Intelligence, as a tool with the potential to positively transform governance and electoral processes, also raise concerns about its ethical use, data manipulation, and reinforcement of existing inequalities.

The evidence produced in the analyses, along with the reflections from the periodic meetings of the Council, indicate pathways and inspiration for the formulation of public policies and advocacy actions. These actions aim to strengthen a democratic environment based on values such as transparency and inclusion, promoting dialogue between different ideological perspectives and combating hostility, toxicity, and disinformation that characterize the current public debate in Brazil.

Addressing these challenges requires broad collaboration between various actors committed to consolidating democracy. It is essential to promote network actions that encourage institutional innovations implemented permanently, with elections and climate policies serving as moments of evaluation and culmination of these initiatives. In this context, coordination between civil society, Big Techs, and governments, especially through institutions like the TSE (Superior Electoral Court) and TREs (Regional Electoral Courts), becomes crucial to strengthening a robust democratic structure adapted to current demands.

Among the priority actions, the enhancement of public policies stands out, with the creation of mechanisms that guide the conduct of candidates, voters, and digital platforms on crucial issues such as climate change and the regulation of Artificial Intelligence. It is essential to strengthen technical devices that promote algorithm transparency, ensure an ethical digital architecture, remove toxic and violent content, and identify illicit practices. The combination of human and algorithmic moderation must be developed in an integrated manner to meet these demands.

Complementarily, it is essential to expand advocacy and education efforts promoted by civil society, with campaigns that not only encourage engagement on climate issues but also empower the population to identify, report, and combat harmful practices amplified by Artificial Intelligence. Pressure on governments and companies must be continuous, aiming for the regulation of AI use in sensitive processes and the increased transparency of digital platforms.

Transparency and access to data circulating on digital platforms remain one of the most urgent demands for researchers, journalists, and observers. Technology companies should provide APIs that guarantee access to detailed information about content flows, contributing to platform accountability and the formulation of evidence-based strategies.

In this way, the following lines of action are proposed:

2.1 Public policies

• Strengthen electoral regulation and oversight mechanisms, such as the continuous and systematic fight against disinformation, hate speech, and gender-based political violence, which can be promoted through partnerships with digital platforms, monitoring protocols, and greater transparency in boosting expenditures;
• Promote civic education and inclusion through the implementation of media literacy programs and incentives for diversity in politics with affirmative policies, as well as formalize the provision of legal and psychological support for female candidates who are victims of violence, either through political parties or through institutions such as regional electoral courts and the TSE;
• Integrate climate policies into campaigns for executive positions and public administrations, establishing climate action plans with clear commitments on sustainability and adaptation to extreme events;
• Combat climate change denialism and disinformation, through partnerships with digital platforms and media outlets, as well as educational campaigns that clarify the gravity and science of climate change;
• Promote environmental equity, prioritizing resources for disaster mitigation and protection of vulnerable populations, especially in contexts of environmental racism;
• Develop comprehensive regulation of Artificial Intelligence for the creation and dissemination of manipulated content in electoral contexts, as well as in non-electoral circumstances;
• Promote digital literacy focused on AI, training citizens to understand its impacts and limitations, especially in political contexts and in crisis situations, such as extreme events related to climate change.

2.2 Digital platform companies

• Improve tools that ensure transparency in digital political campaigns, such as identifying paid content and detailing campaign responsibilities, with effective regulation to prevent the circulation of hate speech and/or disinformation. Combat electoral disinformation through automatic fact-checking systems, partnerships with local fact-checking agencies, and expanded reporting mechanisms for false content, with quick and clear responses to users;
• Prevent algorithmic manipulation by restricting the use of bots for campaigns, requiring strict authentication, and making algorithmic recommendation criteria public during election periods;
• Provide access to APIs and user-friendly interfaces, so ordinary users, research institutions, journalists, and election observers can access and audit the messages circulating on platforms. Create a working group funded by platforms, composed of different representatives from civil society, government, and electoral campaigns, to ensure real-time content monitoring;
• Regarding climate change, expand partnerships with experts and NGOs to review content that spreads environmental disinformation, introducing alerts on posts with false data about climate change and environmental crises. Adjust algorithms to prioritize content from trusted sources, promoting scientific data and informed debates;
• Offer free spaces for NGOs and social movements to run educational campaigns that reinforce climate justice and encourage sustainable practices, increasing the visibility and reach of these initiatives;
• Incorporate technologies to detect deepfakes and visual manipulations in real time, banning the use of AI to create fakes that undermine elections or debates on climate change and extreme events;
• Ensure independent audits of algorithms to prevent the amplification of hate speech and biases, as well as provide clear explanations to users about how content recommendations work;
• Promote the use of AI for the common good, developing collaborative initiatives to map disinformation and monitor environmental crises, with transparency and open access to data.

2.3 Civil society organizations

• Create fact-checking coalitions focused on local and regional narratives, as well as offer media literacy workshops for groups and communities vulnerable to disinformation;
• Develop platforms to monitor electoral spending on social media and publish reports on the circulation of disinformation, analyzing its impacts on public debate;
• Utilize technology to map environmental disasters and crimes such as deforestation, sharing data with the press and regulatory bodies to pressure and hold companies accountable;
• Strengthen climate justice movements led by indigenous and quilombola populations, promoting the formation of international networks that share strategies and experiences;
• Create independent observatories to monitor the use of AI in decisions impacting human rights, and expand partnerships with universities and experts to audit AI systems;
• Advocate for legislation that bans the use of AI for politically motivated deepfakes or disinformation, promoting regulations that ensure transparency in algorithmic decisions and prevent discriminatory biases.

3. Retrospective

Monitoring

Throughout 2024, 41 studies were developed on themes such as public security, indigenous issues, the environment, health, elections, international politics, gender, and communication challenges. During this period, key studies focused on the communication strategies in the race for the mayorships of Brazilian capitals, the debate on the climate tragedy in Rio Grande do Sul, and episodes of violence suffered by marginalized groups, such as indigenous populations, women, and transgender individuals.

Fact-Checking

From the monitoring of social media, Lupa carried out 33 verifications of disinformation circulating on platforms. In the context of this year, the municipal elections, the Olympics, and extreme climate events such as the heavy rains in Rio Grande do Sul sparked waves of disinformation on social media. Notable fact-checks included the claim that judoka Beatriz Souza made the 'L' gesture at the Paris Olympics, the taxation of Brazilian Olympic medals by the government, the claim of unrestricted abortion legalization by Lula, the government's sponsorship of Madonna's concert instead of providing funds to the victims in Rio Grande do Sul, among others. 

Events

Throughout 2024, the project mobilized public authorities and civil society representatives in organizing events addressing topics such as preparations for the G20 in Brazil, the impacts of Artificial Intelligence on democracy, gender-based political violence, and discussions on the municipal elections in Brazil. Among these, the “International Seminar – Artificial Intelligence, Democracy, and Elections” stood out. Held in partnership with the TSE, it featured the participation of Minister Alexandre de Moraes and TSE President, Minister Cármen Lúcia. The event, which took place at the TSE headquarters in Brasília, brought together authorities, experts, and representatives from social media platforms to discuss the impacts of Artificial Intelligence and fake news on electoral integrity. With support from national and international partners, the seminar addressed ethical and social issues related to the use of technology, emphasizing the importance of protecting democracy in the digital age. 

Debate on AI and elections on Jornal Nacional

In August of this year, Fundação Getúlio Vargas analyzed the impact of Artificial Intelligence on elections around the world for Jornal Nacional. In several countries, this technology has been used to spread false information. For more details on the study, access the full report: “Artificial Intelligence is Used in the Spread of False Information; FGV Analyzes the Impact of Technology on Elections Worldwide.”.