The Election

Jorge Santos (fictitious name) knew it was all over. While, in front of him, his advisors and everyone working in his electoral committee frantically discussed, trying to find a solution, Jorge was calm, almost in a trance, feeling the lightness of someone who knows nothing more can be done, his body progressively relaxing as the adrenaline running through his veins was gradually eliminated. "This must be what it feels like when you're dying," Jorge thought, with a sense of detachment from reality, feeling light and almost floating, his thoughts a bit hazy and distant.

Jorge was in the second round of the presidential election, an underdog who, at the beginning of the campaign, was ridiculed by his opponents. However, little by little, his populist speeches, his outsider position, and his pro-family stances captivated voters, along with his charisma and tireless energy. What seemed impossible had happened. Jorge had real chances of being elected, against the center-left candidate, who had started the electoral race with over 40% of the vote intentions and who, despite also being in the second round, had progressively dwindled over the weeks and now barely reached 35% in the second round.

Of course, there was a lot of work behind this turnaround. Through massive and, simultaneously, precise use of digital channels, Jorge and his messages managed to infiltrate the most restricted groups on WhatsApp, Messenger, and Telegram. His Instagram posts, oscillating between sharp criticisms of the system and memes ridiculing his opponent, became so popular that they appeared on open-aired TV. Jorge knew he hadn't always played fair, sometimes distorting reality, but after all, anyone who wants to win a major election has to be aggressive. No one can get elected anymore just by discoursing on their noble intentions. You have to hit the opponent, not let him breathe, maybe even skew the facts a bit until the fight is over. This Jorge knew, despite not being a career politician. But that was the game, wasn't it?

The problem was that the repeated jabs he had thrown and his rise in the polls had generated a powerful counterattack, a left hook that knocked him out. Yes, he knew he had been knocked out, despite everyone around him not having thrown in the towel yet.

The video in which he appeared receiving a bribe to choose a specific supplier for his company was absolutely believable. This internal competition really took place, the winning company was indeed the one that handed him a suitcase full of money in front of the cameras. He really had had that private meeting, everything was perfect and in its place. Except for the fact that this suitcase of money never existed. It had been a trap. He had accepted that meeting just to formalize the exchange of signed contracts, something quick and without major consequences. As Operations Director, he dealt with contracts even much larger than this one. But, somehow, the meeting was recorded. Which wouldn't be a problem, except for the perfect manipulation of the original video. The contracts turned into a suitcase of money. A few compromising phrases were included in the video, with his perfect and clear voice, unmistakable, even his closed "r", typical of the region where he was born. If he didn't know it was all a deepfake, Jorge would also think the video was real.

"Jorge, Jorge!!" – shouted his advisor, trying to bring him out of his near-absent state. "We will inspect the material, and surely, we will have technical evidence that it is a fake video. Perfect, but fake!".

Jorge shrugged. He knew that there was still no certain way to detect AI-manipulated videos. There were several applications that claimed to detect material produced with this technology, but the error rate was still very high. The tools worked by analyzing subtle inconsistencies in the videos, like small lighting failures or lip-sync errors, but these flaws were becoming increasingly difficult to detect as the technology advanced.

His advisor continued talking about the inspection, mentioning that one of the most promising tools for detecting deepfakes analyzed facial microexpressions – quick and involuntary muscle movements in the face that were almost impossible to replicate perfectly. However, Jorge knew that even this tool could be fooled by high-quality deepfakes, especially if created by specialists.

There were several applications that claimed to detect material produced with this technology, but the error rate was still very high. Jorge knew that, for example, a popular deepfake detection software often produced false positives – accusing legitimate videos of being fakes – or false negatives, letting manipulated videos pass. Only when commercial tools that included some sort of watermark were used was the detection unequivocal. Ha, ha, ha... laughing was the only option. Any forger, when enough money was involved, would obviously use DeepWeb tools. And it wouldn't even be very expensive or time-consuming. Nowadays, even a good amateur could create imperceptible videos, audios, and texts.

"Jorge, since the video is anonymous and doesn't carry any trace or indication of being generated by AI, we'll have to work with Internet providers to find the upload point. Using digital forensic analysis techniques to examine metadata and compression patterns, we might be able to follow the traces and get to the author of this hoax," said the head of cybersecurity for his electoral committee.

"Alright, let's move forward," agreed Jorge. He knew he would deny everything and point out the falsification. The problem was that there was less than a week left until the election, and as Jorge had well learned in his 60 years of life, well-crafted lies spread much faster than the truth.

Jorge had never been a saint, and he knew that very well. But the blow he received was a low one. And fatal. He knew his case was lost. But, from now on, he would use the more than 40% of valid votes he received in the first round to lead efforts to fight against this phenomenon of public opinion manipulation, which threatened to destroy the legitimacy of future elections. He didn't know if it would work, probably not. But it would be worth trying.