As the year draws to a close and we charge ahead into 2025, Meedan is looking back on a record-breaking year for elections worldwide. In 2024, across more than 60 countries, more people voted in general and legislative elections than ever before in world history. In nearly every one of these contests, digital ecosystems played a pivotal role in determining what information voters accessed and how they engaged with that information.

At Meedan, we are firmly committed to the idea that trustworthy media — from hard-hitting investigative news stories to single-issue fact-checks to tip sheets on voting rights — form critical underpinnings of a healthy, democratic society. In 2024, we doubled down on this commitment by facilitating election-focused coalitions in 11 countries across five continents. Check, our open-source software platform, enabled 53 partner organizations to collaborate with each other and connect with their audiences on messaging services like  WhatsApp and Messenger. Users could send in questions about rumors they’d seen online and receive trusted responses in return.

Our 2024 election projects ranged from large-scale initiatives reaching tens of millions of voters in countries like Mexico and Indonesia to narrowly scoped projects centered on specific issues like gender-based violence in places like Pakistan and Togo.

How AI did — and did not — affect elections

AI did not upend elections as some predicted in early 2024, but it certainly made its mark. In Pakistan, AI-generated video and voice recordings played a major role in the success of Imran Khan’s Pakistan Tehreek-e-Insaf (PTI) party. There, and in both Brazil and the U.S., women and LGBTQ+ candidates were top targets of AI-generated deepfake images intended to sexualize or associate them with scandal.

But in other cases, instead of being used to denigrate minority candidates or to trick voters into believing things that weren’t true, AI was used to appeal to voters’ emotions by casting politicians in a flattering light.

A Prabowo campaign poster in Jakarta. Photo by Kaliper1, CC BY 4.0.

With the help of AI, Indonesia’s former Defense Minister Prabowo Subianto — an ex-military leader who has been accused of facilitating human rights abuses — rebranded himself as a cuddly grandpa figure. A cartoonish rendering of the candidate and his running mate, generated by MidJourney AI, became the winning campaign’s signature image.  

In the U.S., after Donald Trump repeated the baseless allegation that Haitian immigrants in Ohio were eating their neighbors’ pets, supporters began posting AI-generated memes in which Trump could be seen cuddling or rescuing kittens and ducks from a vague specter of harm. The incredible popularity of these images on social media underscored the degree to which AI can shape how voters perceive and connect with political candidates.

Low-tech tactics undercut voter confidence

While deepfakes and AI-generated media certainly made the rounds, social media influencers and regular internet users alike proved that once the seeds of uncertainty have been sown, time-tested tactics remain effective for ginning up suspicion about electoral processes.

Across the globe, we saw trends of distrust in everything from the administration of election-related laws to the conduct of poll workers and the status of official pens used for filling out ballots. In Mexico, one prominent rumor held that these pens contained erasable ink. From Brazil to Mongolia and the U.S., rumors of voting machines malfunctioning — or being deliberately programmed to ensure certain outcomes — were pervasive on social media.

Photo by: ProtoplasmaKid, CC BY-SA 4.0. This image has been adapted.

Gender on the ballot

Some of our partners had also anticipated — well ahead of major elections — that gendered stereotyping, harassment, and hate speech would be deliberately used to target a wide variety of candidates. Partners on three continents documented numerous cases of candidates being criticized on the basis of their gender identities and those of their associates. In Pakistan, Togo, and Argentina (which held elections in late 2023), gendered narratives were deployed to denigrate political candidates and their families, illustrating that this concerning trend is indeed global in nature.


Pakistan

Ahead of Pakistan’s parliamentary elections in February, AI-generated voice clone audio files of jailed former Prime Minister Imran Khan took social media by storm, dovetailing with a robust digital campaign on the part of the PTI party, of which Khan is the de facto leader. In an election where 60.8 million people exercised their right to vote, PTI claimed a majority of seats in parliament despite significant roadblocks, including social media platform shutdowns during major party events.

In previous elections, our partners at the Digital Rights Foundation (DRF) had observed waves of misogynistic harassment and gendered disinformation targeting candidates for office. This time around, they decided to focus on how these kinds of messages were shaping online discourse. A top target of gendered disinformation they identified was Maryam Nawaz, a member of the provincial assembly in Punjab, who stood and won election as Chief Minister of the province in February. The daughter of former Prime Minister Nawaz Sharif, Maryam Nawaz is a prominent member of the Pakistan Muslim League-Nawaz party, which has close ties to the country’s military. DRF logged numerous social media posts claiming that Maryam Nawaz had “seduced,” “charmed,” and “enticed” Pakistan’s army to go against Khan.

While attacks on female leaders in politics were nothing new, anti-transgender narratives are one area of Pakistani politics in which harassment seems to have increased. Evidence of this trend was clear in the case of Bilawal Bhutto of the Pakistan People’s Party, who sought and won reelection to the National Assembly. “He was attacked due to being too feminine,” said Seerat Khan, a program manager at DRF. People attempted to malign Bhutto by suggesting that he is transgender. 

When politicians are labeled as transgender, Khan explained, “they aren't taken as seriously.” As recognition of trans people’s rights continues to gain ground in Pakistan, Khan says her team at DRF expects anti-trans harassment to increase too.

Photo by Wotancito, CC BY 4.0. This image has been adapted.

Mexico

In June, voters made history when they elected Claudia Sheinbaum as Mexico’s first female president. To support the dissemination of trustworthy information about the election, Meedan partnered with the National Electoral Institute (INE) — Mexico’s highest election authority — and worked alongside Agence France-Presse, Animal Político, and Telemundo.

What did Mexican voters want to know? Three of the most common questions we received referenced rumors about problems within the country’s electoral systems. This is nothing new, according to Alfonso Peralta, who serves as deputy director for information monitoring and analysis at INE. Peralta explained that public distrust in Mexico’s election system routinely surfaces during electoral contests. But the ways in which these narratives travel are changing.

“In this election, we observed that WhatsApp was a platform with a high volume of disinformation,” Peralta said. “It was really important for us to be able to rely on a tool like a chatbot that helped us to channel citizens’ questions.”

Brazil

In October, voters in more than 5,000 cities throughout Brazil cast ballots for mayoral positions and other local offices. In the lead-up to these elections, Meedan teamed up with Pacto pela Democracia, a coordinating body for a coalition of more than 200 civil society organizations working to increase citizens’ access to reliable information.

Although these were local contests, popular social media narratives made it clear that the country’s deeply polarized national political paradigm had cast a long shadow. Most verification requests that the coalition received reflected public distrust in the country’s electoral processes — an increasingly prevalent issue since the administration of former President Jair Bolsonaro. Rumors about faulty voting machines and the Supreme Court interfering in election processes — which dominated online discourse during Bolsonaro’s unsuccessful 2022 reelection bid — resurfaced with force, proving that some disinformation dies hard.

More than half of the queries we received reflected concerns about the integrity of Brazil’s electoral systems

The road ahead

Over the course of the year, our election work enabled us to gain new insights into how information moves across digital spaces and how it is disseminated, distorted, and sometimes even disappeared from our social feeds and messaging groups. With this summary report, we’ve allowed ourselves to take a long look at what we learned throughout 2024 so we could share new insights with our colleagues and consider how to build upon this work as we strive to increase our impact in 2025. We look forward to exploring these considerations as a team, and as a community, in the months to come.

We collaborated with 53 partner organizations worldwide to design and carry out our 2024 elections projects. We extend special gratitude to our lead partners in Brazil, Mexico and Pakistan, whose work we highlight in this essay.

Pacto pela democraciaINE MexicoDigital Rights Foundation

The 2024 elections projects featured in here would not have been possible without the generous support of these funders.

SkollSIDAPatrick J McGovernSVRI
Tags
Elections
Election integrity
Election Misinformation
Misinformation
Mexico
Generative AI
Pakistan
Brazil
Disinformation
Gendered disinformation

Footnotes

References

Authors

Words by

No items found.
No items found.
Words by
Organization

Published on

December 13, 2024
December 9, 2024