deepfakes are in the viewfinder as the legislative elections approach

In South Korea, as elsewhere, a vote is planned for 2024, for legislative elections. Faced with the explosion in the use of AI in electoral campaigns, the commission which oversees elections has just intervened to put limits on its use.

Published


Reading time: 2 min

"AI Yoon"the digital avatar of current President Yoon Suk-yeol, with the team that developed it, on February 7, 2022. This avatar was used in the campaign to make it "cool" with young people.  (JUNG YEON-JE / AFP)

During the year 2024, more than 4 billion people will be asked to vote. That’s one in two Earthlings! There are major elections in the United States and Europe, but also many national elections in Asia. For their campaigns, candidates are very tempted to use artificial intelligence, which has made enormous progress.

In South Korea, the electoral commission was forced to issue new regulations this week to prevent the explosion of so-called deepfakes on social networks. These fake videos or audios, created with the help of artificial intelligence, are increasingly used to deceptively convey messages.

Regulate deepfakes, which mainly target young people

By entering the program with real videos and the sound of a personality’s voice, they find themselves declaring exactly what we want with disturbing veracity. This can make it possible to discredit a political opponent or, on the contrary, to highlight a candidate.

In South Korea, several elected officials have already started using these deepfakes in the campaign for the legislative elections in April. And during the 2022 presidential election, the current conservative president, Yoon Suk-yeol, himself used his image extensively in fake videos to seduce young people on social networks.

The ban decided this week by the Korean electoral commission therefore targets the use of videos, images or even sounds, created by artificial intelligence to support a candidate or to denigrate another.

Ban or partial authorization depending on the country

The commission explains that these creations interfere too much with the democratic process. It does not completely ban deepfakes. Indeed, these videos are still authorized if they simply serve to report in a neutral manner on the activities of a political party.

Many other countries are also worried and are starting to regulate these deepfakes in politics. Almost everywhere, regulators find themselves forced to react urgently. Since the start of the year, 14 states in the United States have taken measures similar to what South Korea is doing. There, while the presidential elections will take place in November, either deepfakes are completely banned in the presidential campaign, or they force the candidates’ teams to explain in their videos that they are fakes, created with artificial intelligence.


source site