'Fake news', plagiarism… NewsGuard warns of explosion of AI-powered news sites

Post it

Reading time: 4 minutes

A computer generated photo of robots working on keyboards.  (ANDRIY ONUFRIYENKO/GETTY IMAGES)

Carbon copies of established media, these disinformation sites proliferate and attract advertisers.

You research a current event, click on a website address, read the article then shut down your computer. Without knowing it, you may have just read a text written by artificial intelligence (AI). AI-powered websites have proliferated in recent months. NewsGuard, a start-up that specializes in monitoring misinformation, is sounding the alarm. In one of its reports, it had identified 49 addresses of such sites in May 2023 ; Today, it lists 767, including 15 in French.

These sites, with little or no human presence behind them, publish thousands of articles every day. Without real human control, they tend to spread false information, NewsGuard explains. In November 2023, the information site Global Village Space announced, for example, that Benjamin Netanyahu had caused the suicide of his psychiatrist. Artificial intelligence was unable to differentiate this joke, published in a satirical newspaper in 2010, from real news, and turned it into article-like content, NewsGuard revealed.

Fake information and catchy headlines to generate clicks

Humans are indeed at the origin of these sites, but their identity remains unknown. “Some have understood the financial interest, advertising generates financial income” explains Chine Labbé, editorial director of NewsGuard. Articles are formatted in such a way that they appear in the first results in search engines. “That's what attracts advertisers; they want their ad to appear, so they target the sites with the best reports for distribution,” notes the journalist.

To maximize their exposure, they publish shocking information. Their specialty : announce the death of living people. So, in January 2024, she is journalist Deborah Vankin, from the American newspaper LA Times, who paid the price. Why ? His name was typed into search engines several times after one of his columns was published. “When AI notices that a name is highly searched for, it's very common to produce obituaries, which is the type of article that people click on.” promotes Chine Labbé, clicking is the number one goal.

“Others post polarizing content to cause mayhem.”

Chine Labbé, editorial director of NewsGuard

at franceinfo

After identifying conflicting views on the web, such as the Israeli-Palestinian conflict, the AI ​​produces articles on those topics to attract readers. The goal remains the same: generating reactions and generating clicks. But then how to identify this content ? Some techniques make it easier to spot them. First, they they are full of repetitions and contradictory elements. “If you ask an artificial intelligence to write a recipe for an omelet with a cow's egg, it can certainly do it.” notes Amélie Cordier, an artificial intelligence engineering expert interviewed by franceinfo. In general, these texts contain very few references and sources. Finally, these sites are super productive. This is the case of Interstars, one of the 15 French-language websites generated by AI. Overproduction, incentive titles and replays : everything is there. So many signs that should make you suspicious.

But sure The sites are more subtle and repeat articles from other online media verbatim. Like L'Observatoire de l'Europe, which NewsGuard spotted in its Reality Check newsletter. This website copies articles from Euronews, an online media outlet affiliated with a European TV channel. The similarities are striking, some articles are carbon copies, down to the letter. The last: an article contradicting the rumor of the death of King Charles III. Originally published on the Euronews website on March 19th, this article is in L'Observatoire de l'Europe on the same day, but this time the author is identified as one Jean Delaunay.

On the left the article published by Euronews.com on 19 March 2024. on the right, the one published on the same date on the Europe Observatory website.  (FRANCE INFORMATION)

The European Observatory is not his first effort. In December 2023, he already reprinted an article from Euronews on ChatGPT and, even more recently, an article about St. Patrick's Day was plagiarized. Each time, the signature appears as “Jean Delaunay”, introduced as the site's founder and professor of political science at several universities. “supposed”. Upon investigation, we found no significant trace of this pseudo-author.

Whether plagiarizing other media or spreading false information, these sites abound. For Chine Labbé, the solution is “Educate companies not to advertise on these types of sites in order to minimize their visibility”. There is currently no legislation on this matter.

Leave a Reply

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