F.D. Flam: It's getting harder to study fake news

Researchers who study misinformation are confronting a new problem: public scorn. And it’s not just in the form of online trolling. These scientists are losing funding, watching their research centers close down, and getting barraged with subpoenas.

Given the rapid changes to news, social media and information sharing, you’d think there’d be more support for studying how people learn about the world. Instead, critics are wrongly conflating their work with censorship.

In the New York Post, for example, a story hammered a group of psychologists as concocting “fake science” to justify censorship. It’s easy to see why their paper, published last week in the journal Nature, hit a nerve. The researchers found that conservatives shared more information from low-quality news sites on social media than liberals did.

While the idea of news quality sounds subjective and prone to bias, the scientists didn’t make that judgment themselves. The researchers asked three groups to weigh in: professional fact checkers, a politically mixed group of laypeople, and a group of Republicans. Each group determined what was a high-quality source (a news organization that mostly gets it right, but can sometimes make mistakes) or a low-quality one (a publisher that tends to make things up out of whole cloth).

After each group determined what counted as low-quality news, the team looked at who typically shared that type of news. Each time, they found that extreme partisans on both sides were more likely to share this misleading content. And each time, those on the far right contributed more garbage to the information effluent.

The study doesn’t justify censorship of conservative views, although it does offer an explanation as to why right-wing social media accounts are more likely to be suspended. It shouldn’t be attacked just because it’s offending people. That flies in the face of the spirit of free inquiry.

Edward Tenner, a historian and lecturer on technology and culture, explained to me that the pushback against the paper could be what’s known as reactance — a tendency for people, when told they’re wrong, to double down. Stirring up antipathy is always going to be an occupational hazard for people who study misinformation, rumors, pseudoscience and quackery.

Adding to that is the problem that many people don’t mind lies — they only abhor lies spread by their political opponents. In a recent article in The Ohio Capital Journal, Minjae Kim, an assistant professor of management at Jones Graduate School of Business at Rice University, called this acceptance of certain lies “moral flexibility.” Citing recent research he collaborated on, he wrote that some supporters of former President Donald Trump recognized that Joe Biden won the 2020 election, but justified Trump’s claims to the contrary because they believed “the political system is illegitimate and stacked against their interests.”

And people on the left didn’t seem upset with Biden’s erroneous 2021 claims that people vaccinated against COVID couldn’t spread the disease to others. That had been the hope, of course, but the Delta variant had already shown that not to be the case. Partisans may have considered Biden’s statement acceptable, though, reasoning that because the vaccines had the power to save lives, the details didn’t matter.

Tenner considers the relevance here of the Italian saying, “Se non è vero, è ben trovato” — even if it is not true, it is a good fabrication, or good story. That’s attributed to 16th century philosopher Giordano Bruno who had some forward-thinking ideas … and was burned at the stake.

The expression might describe the way JD Vance reacted to Trump’s statement that immigrants in Ohio were eating cats and dogs. There were no documented incidents of such activity, but Vance attempted to justify the rumor by saying that it called attention to problems surrounding immigration.

These kinds of wild stories are the very type social media algorithms tend to amplify with the help of foreign intelligence agencies and automated “bot” accounts. Yet there are ways to moderate the information stream other than taking things down.

“Part of the research that we’re doing right now is to develop models so that we can evaluate intended and unintended consequences of different moderation schemes,” said Filippo Menczer, a professor of informatics and computer science at Indiana University. But he said these efforts have gotten much harder due to political attacks. Meta and X have also restricted data access to many researchers.

Instead of deleting posts or “shadow banning” users who express non-mainstream views, social media companies can fight incorrect or disputed information with additional information. In dynamic areas such as science and medicine, moderation should be transparent because fact-checkers have mistaken legitimate minority opinion and insightful dissent for misinformation. In some tests, the crowdsourcing-based “community notes” feature on X helped diffuse medical misinformation.

Picking fights with scientists won’t make our information problems disappear. A more consistent view for those who are pro-free speech and anti-censorship would be to embrace free inquiry into our information ecosystems — and to applaud those who scrutinize the algorithms that influence what we think and how we vote.

_____

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

F.D. Flam is a Bloomberg Opinion columnist covering science. She is host of the “Follow the Science” podcast.

_____