The threat of deepfakes looms over polling stations in the United States

Sign up for the International newsletter What seems far away matters more than ever

Sign up for it

Last week, a photo of Donald Trump surrounded by black boys was circulating on the networks. The image was an attempt to arouse sympathy among the voters of this community. It was also fake.

THE fake news and the deepfakes (AI-generated images and videos) have been circulating the internet for some time now. In the last election of 2020, Facebook already reported an increase in disinformation during the campaign. But this year they pose an even greater threat in a polarized society in which many citizens remain convinced that the 2020 elections were a fraud.

The rapid development of AI makes it possible to create increasingly realistic videos and photos, which threaten to further fuel polarization during the campaign. “These tools are becoming more accessible and cheaper for the general public. It’s becoming easier for anyone to create this kind of content to influence public opinion,” explains McKenzie Sadeghi, publisher of NewsGuard, a company specializing in detecting and deny deepfakes i fake news.

AI is within everyone’s reach. It is not only the case of creative citizens deepfakes to favor their candidate (as was the case with photographs of Trump with black voters), but also the parties themselves have used the tool as a weapon to attack the opponent. Last June, Republican candidate Ron DeSantis’ campaign team released a video featuring AI-generated footage of Trump kissing and hugging Dr. Anthony Fauci, an unpopular figure in the country following his management of the pandemic.

The threat of deepfakes it extends beyond the country’s borders. In February, US intelligence released a report warning that Russia, China and Iran will try to influence public opinion in the run-up to the election. “It’s a fact that it’s already happening. A few weeks ago, an AI-made image of Biden dressed as a military man was circulating and people were sharing it warning that there was an escalation in the war between Israel and Hamas. We detected that many of the media Russians shared and amplified this deepfake. In total, we counted 700 pages where it appeared,” details Sadeghi.

The vote, in the center of the target

The moment an image (whether real or not) hits the internet it becomes uncontrollable. The speed with which the deepfakes they can spread is overwhelming. “It only takes twenty seconds to decide whether to retweet or not. Nobody can read an article in twenty seconds, much less compare it,” explains Nathan Walker, professor at Northwestern University and director of the Center for the Psychology of Media and Influence Social.

But the battle seems lost. “It’s very difficult to track down all the deepfakes that there is, especially considering that many of the videos are shared privately through messaging apps. We can’t even follow them and refute them,” says Carah Ong Whaley, PhD in American politics at the University of Virginia and member of the non-partisan platform Issue One.

One of the great fears of the elections is that, beyond that with the deepfakes attempts are made to change the meaning of the vote, confusion is also created during the election day. “Let’s just in case someone spills one deepfake saying that the polls have already closed and that means that a lot of people don’t go to vote,” says Sadeghi. It’s a scenario that Whaley sees as very possible: “I think we’re going to find a lot of deepfakes that they will try to prevent people from going to vote in their school or directly saying that they can no longer go there. It’s very dangerous and I think it will be one of the biggest challenges of the election.”

Trust in institutions, threatened

The deepfakes they fall on a long-tilled ground. “The 2020 election had as much transparency as possible. And there are recordings of Trump and some of his supporters calling state legislators encouraging them to influence the vote count. There was no deepfakes. They were real people, politicians, saying the election result was fake. This is worse than any deepfake“, explains Walter. “The big underlying problem is that trust in institutions has been lost,” points out the communication professor. Whaley agrees: “I have a focus group of Republican voters, all of them have higher education, but the their big concern in this election is voting by mail. They don’t trust it. If you don’t trust the voting system, there is no more. There is a risk that [amb els deepfakes] trust in the system ends up being completely destroyed.”

Walter sees politicians’ preoccupation with AI as a “turkey’s head” for the main problem in American society: polarization and mistrust of institutions. “There are politicians who want to ban TikTok, but it won’t work. We need technological solutions, but not only. We need to rethink our education and our economy. As long as the economy continues to be based around clicks andengagementall the incentives that exist make it easier to give wings to disinformation,” says the Northwestern University professor.

2024-03-16 19:59:02
#threat #deepfakes #looms #polling #stations #United #States

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *