Scientific journal articles vulnerable to bots

Scientific journals are easy targets of automated software that post links to social media, often with misinformation, according to a study published by JAMA Internal Medicine.

Why it matters: Automated disinformation campaigns that harness legitimate scientific research could further erode the public's understanding and trust in science, particularly around COVID-19.

Driving the news: Researchers looked at 563 Facebook groups that shared links to a Danish trial, published in the Annals of Internal Medicine in November 2020, that supported mask-wearing.

• About 39% of the posts that provided direct links to the study were in Facebook groups most affected by automation while about 9% of the Facebook groups were least affected by automation.
• Among posts made to groups most affected by automation, about 20% claimed masks harmed the wearer and 51% made conspiratorial claims about the trial. In comparison, among groups with the least automation, 9% claimed masks harmed the wearer and 20% made conspiratorial claims about the trial.

The researchers said their findings indicated the study was likely the subject of a campaign to disseminate misinformation.

• They recommended legislation to strengthen penalties for those behind automation, greater enforcement of rules by social media companies and counter-campaigns by health experts.