American TV presenters, though, are stars at artificial intelligence trickery

Videos viewed thousands of times on Facebook extoll the benefits of slimming or diabetes products by exploiting the image of star TV presenters on US channels, such as CBS or CNN.

But these videos are entirely made by Artificial Intelligence (AI) software. the Deep fakesThese increasingly frequent and realistic digital manipulations are putting the reputation of traditional media at risk.

Some broadcasters and journalists whose identities were stolen interacted directly with the broadcast of these manipulated videos on networks.

“I have never heard of or used this product! Don’t be fooled by these AI-generated clips,” CBS anchor Gayle King posted on her Instagram page in October.

Other videos even use and distort billionaire Tesla CEO Elon Musk’s words for commercial purposes.

these Deep fakeswhich praises all kinds of products and questionable investment schemes, often refers to e-commerce platforms and ephemeral websites, which disappear a few days after being distributed on networks.

Since 2020, Meta – the parent company of Facebook and Instagram – has banned the distribution of such videos on its platforms, with the exception of parodies and some satirical content. However, these clips, several examples of which have been analyzed and verified by AFP, are still circulating freely online.

Sound reproduction

“There is a resurgence of these types of videos using just a two-minute audio sample and reproducing someone’s voice in a completely new imaginary sequence, with synchronized mouth movements,” explains Hani Farid, a professor specializing in digital technology at the University of California. Berkeley.

Audio-visual characters, because of their constant on-screen presence, are easy targets for training AI software.

This is a troubling trend because the public has created a bond of familiarity with these impersonated public figures, said Andrea Hickerson, dean of the University of Mississippi School of Journalism.

“It’s really dangerous because people don’t expect misinformation to be materialized in this way,” she told AFP.

The content is then presented “as it is in traditional media”.

“A crisis of confidence”

AI-manipulated content is also playing an increasing role in financial fraud, which has cost Americans about $3.8 billion in 2022, according to FTC calculations.

These scams have reportedly targeted people in several countries, including Canada and Australia, and have cost some people tens, if not hundreds, of thousands of dollars.

“Scams are becoming increasingly sophisticated as criminals combine traditional fraud tactics with scams involving cryptocurrencies and AI software,” attorney Chase Carlson said in a blog post published earlier this year.

Americans are also increasingly concerned about the use of artificial intelligence, especially in politics.

More than 50% of Americans expect that lies generated by artificial intelligence will have consequences in the 2024 presidential election, according to a poll published in September by the media outlet Axios and Morning Consult, an economic research firm.

Agence France-Presse has already analyzed manipulated video clips in which US President Joe Biden appears announcing a general mobilization or in which former Secretary of State Hillary Clinton announces her support for Republican Florida Governor Ron DeSantis in the upcoming presidential elections.

According to an October Gallup poll, only a third of Americans trust news media “a lot” or “somewhat,” matching the lowest level recorded in 2016.

The spread of this content, which can sometimes be easily spotted due to its poor quality, therefore poses a risk of fueling a “crisis of confidence” in the media and institutions on the part of the public, according to Rebecca Tromble, director of the institute. For Data, Democracy, and Politics at George Washington University.

“High-quality information is always out there, and with a large dose of skepticism, we can separate fact from fiction,” stresses the expert, who urges caution before transmitting any type of content online.