
Alex Eben Meyer
A growing number of websites mimic the look of traditional news outlets and purport to be produced by journalists who are grounded in fact. But they often aren’t. Much of the content is created by artificial intelligence (AI) and promotes a political agenda. What do readers think? A new study led by Kevin DeLuca, an assistant professor of political science at Yale, finds that almost half of readers actually prefer these algorithmically generated sites over traditional ones.
The sites take on the look and feel of legitimate local newspapers—even adopting names like the Detroit City Wire or North Boston News. But their stories are laced with partisan messaging and are often funded by political campaigns. Unlike outright fake news sites, many of their stories draw from real press releases, but they also publish partisan content that suits their ends. The phenomenon has been called “pink-slime journalism.”
DeLuca, whose research focuses on elections and media, wanted to know if people could tell the difference between human-produced and algorithmically produced news sites. The researchers were also interested in identifying whether a media literacy intervention could help. In the study, participants were shown both types of websites, with some participants receiving a “tip sheet” detailing how to spot credible sources by checking bylines and finding information about the news site.
The researchers found that the intervention worked as intended: Participants who received the tip sheet used the cues. But it didn’t necessarily change what they preferred. Even after reading the tip sheet, 41 percent still favored the algorithmically generated site, down from 46 percent in those who hadn’t.
The findings suggest that convenience, familiarity, and confirmation of readers’ existing beliefs often outweigh concerns about authenticity. Pink-slime outlets also may be more appealing than real news sites because they generally don’t have to include intrusive ads that can disturb the reading experience.