A team of researchers has developed a test to measure people's susceptibility to misinformation, finding that on average, US adults fail to classify one-third of real or fake headlines into their correct category.
The test, which is takeable online, asks the user to rate 20 headlines as real or fake. Ten of the headlines were real, while the other 10 were generated by artificial intelligence (AI) chatbot ChatGPT, before being vetted and selected by a panel of experts.
When refining and selecting the fake questions, the team looked for common features of misinformation, such as appealing to people's political biases, and tinges of conspiracy theories. The real headlines were selected from the Media Bias Fact Check database, chosen from news sources that ranked highly for factual reporting and low on bias.
The team, conducting the research on 1,516 US adults, found that the average American got 13 out of 20 headlines correct.
The headline people got correct the most was "Republicans Divided in Views of Trump’s Conduct, Democrats Are Broadly Critical", with 80 percent of participants rating it as real. Younger participants scored lower on the misinformation test, getting on average 12 out of 20 to older adults' 15.
You can take the test yourself.
The findings are published in Behavior Research Methods.