A collage on how a “deepfake” is detected / Web
With the United States presidential elections around the corner, on November 5, fear among experts has focused on an invisible but powerful enemy: generative artificial intelligence (AI). This technology, which allows the creation of manipulated images, videos and audios, has established itself as a potentially dangerous tool for the integrity of the electoral process, to the point that many voters may not distinguish between reality and fiction in the avalanche of content that circulates. on social networks.
However, what has filled the feeds of the networks are not the sophisticated deepfake videos that were feared, but much more absurd and cartoonish content. Among the most viral examples are a video of former President Donald Trump riding a cat while wielding a rifle, or Vice President Kamala Harris with a mustache and dressed like a communist. These ridiculous images have left no one indifferent, and despite their apparent humor, they raise serious concerns.
The tangled campaign
As the US presidential campaign progresses, AI-generated images have become a common resource, circulating online with alarming regularity. Although many of them are so absurd that it seems impossible to take them seriously, their impact goes beyond the comical. Even the most ridiculous memes can serve to spread false or, worse yet, racist and xenophobic messages. Candidates and their followers share them openly, using them as weapons in the digital battle.
Trump supporters have spread a large number of AI-generated memes
Trump and his allies, for example, have been especially prolific in spreading AI-generated memes. One of them, posted on his Truth Social network, showed him surrounded by cats and ducks on a private jet, while another depicted a group of kittens holding a sign that read: “Don’t let them eat us, Vote for Trump!” . These images, although comical in appearance, reinforced a false xenophobic narrative about Haitian migrants, fueling racism and generating real threats against vulnerable communities.
The impact on Brexit
Christopher Wylie, the man with pink hair, surprised the world by uncovering the key role that Cambridge Analytica (CA) played in the result of the Brexit referendum. When asked if the company’s intervention was decisive in the victory of the United Kingdom’s departure from the European Union, he responded without hesitation: “Absolutely.” With that statement, he revealed how CA, through data illegally obtained from Facebook, would have influenced elections as momentous as Brexit and the elections that brought Donald Trump to the presidency of the United States.
The use of networks to influence electoral behavior is not new
Wylie, former director of research at CA, explained in an interview with several European media that the Brexit vote was won by less than 2% of the votes and that the company spent million-dollar sums on highly personalized advertising campaigns, aimed at small groups of voters. These efforts would have been enough to tip the balance in favor of leaving the European Union. According to Wylie, CA did not act directly, but through an allied company, AggregateIQ (AIQ), hired by the Brexiteers.
What CA did was simple but effective: Using personal data collected from Facebook, the company profiled users based on their interests, behaviors, and opinions. With that information, he designed advertising messages tailored to each group, seeking to reinforce prejudices or previous beliefs. The campaigns fed on fake news and misinformation, and also circumvented legal limits on electoral spending, which has sparked investigations in the United Kingdom.
The use of social media to influence voting behavior is not new, but CA took it to another level. By targeting personalized messages to specific audiences, and using bots and fake accounts to amplify the reach of those messages, the consultancy managed to create a feeling of massive support that distorted reality.
