If you’ve read a recent social media posting that offers some questionable advice or dubious news about COVID-19, you’re not alone. In fact, users on Twitter have, on average, about a 50 percent chance of reading information regarding the disease that most likely came from a bot, according to findings from a recent study.
Researchers from Carnegie Mellon University delved into 200 million tweets that have been authored since January about coronavirus. Of those examined, 45 percent were issued out by accounts that behaved more like “bots” than human users of the social media site.
Bots are “automated user accounts that interact with Twitter using an application programming interface” to tweet, retweet and even message human users online, according to consumer software company Symantec. They can sometimes be used for good reasons — such as providing information on weather or earthquakes to relevant users — but they can be used for iniquitous purposes, too, like interfering in the political process or waging disinformation campaigns on topics their developers want to promote. Some research suggests that as many as 15 percent of all users on Twitter are actually bots.
The team of Carnegie Mellon researchers said it was too soon to say who was behind the bots’ disinformation campaign, but the content in the tweets themselves seem to suggest the motivation was to create division among American users on Twitter — similar to what happened during the 2016 presidential campaign.
The bots appear to have disseminated at least 100 different false storylines about coronavirus that are being spread online, the researchers concluded. These include conspiracy theories, like that hospitals are putting mannequins in hospital beds to make the crisis appear more damaging than it is, or that 5G cell towers spread the disease, both claims that are patently false.
According to the researchers, the bot activity is two times higher than what they predicted it would be, based on previous analyses of bot behavior during past disasters, crises and elections.
In February, a separate analysis of activity from Bot Sentinel found that bots were spreading a variety of false claims about the disease, including that China intentionally created the virus or was trying to weaponize it before it spread, or that Democrats were overstating the concern about COVID-19 in order to make President Donald Trump look bad.
The Bot Sentinel analysis also concluded that most of the tweets from coronavirus-based bots were favorable to the president.