ILLUSTRATION: TOMASZ WALENTA

Tech

Detecting Fake News Takes Time

par Featured articles licensed from The Wall Street Journal | le 28 February 2020


 

 

 

A new study shows that people can tell when a story on social media is bogus—if they take a moment to think.

By

Alison Gopnik – The Wall Street Journal

A few weeks ago, I took part in a free-wheeling annual gathering of social scientists from the academic and tech worlds. The psychologists and political scientists, data analysts and sociologists at Social Science Foo Camp, held in Menlo Park, Calif., were preoccupied with one problem in particular: With an election looming, what can we do about the spread of misinformation and fake news, especially on social media?

Fact-checking all the billions of stories on social media is obviously impractical. It may not be effective either. Earlier studies have shown an ”illusory truth” effect: Repeating a story, even if you say that it’s false, may actually make people more likely to remember it as true. Maybe, in our highly polarized world, they can’t even tell the difference; all that matters is whether the story supports your politics.

But new research contradicts this pessimistic picture. David Rand of MIT and Gordon Pennycook of the University of Regina have suggested that “cognitive laziness” may be a bigger problem than bias. It’s not that people can’t tell or don’t care whether a story is true; it’s just that they don’t put in the effort to find out.

A new study in the Journal of Experimental Psychology by Profs. Rand and Pennycook, with Bence Bago of the University of Toulouse, shows that if you give people time to think, they do better at judging whether news stories on social media are true or false.

The researchers showed more than 1,000 people examples of true and false headlines that had actually appeared online—real fake news, as it were. Some headlines were slanted toward Republicans, like “Obama was going to Castro’s funeral until Trump told him this,” while others were slanted toward Democrats, like “Gorsuch started ‘fascism forever’ club at elite prep school.”

The researchers asked participants to judge whether the headlines were accurate. One group was allowed to take as much time as they wanted to make a judgment, while another group had to decide in seven seconds, while they were also trying to remember a pattern of dots shown on the screen. Then they had a chance to think it over and try again. The participants also filled out a questionnaire about their political views.

As you might expect, people were somewhat more likely to believe fake news that fit their ideological leanings. But regardless of their politics, people were more likely to spot the difference between real and fake news when they had time to think than when they had to decide quickly.

Of course, when we browse Twitter or Facebook, we are more likely to be rushed and distracted than patiently reflective. Lots of items are pouring quickly through our feeds, and nobody is asking us to pause and think about whether those items are accurate.

But it would be relatively easy for the platforms to slow us down a little and make us more thoughtful. For example, you could simply ask people to rate how accurate a story is before they share it. In preliminary, still unpublished work, Profs. Rand and Pennycook found that asking people to judge the accuracy of one story on Twitter made them less likely to share others that were inaccurate.

Cognitive science tells us that people are stupider than we think in some ways and smarter in others. The challenge is to design media that support our cognitive strengths instead of exploiting our weaknesses.