American researchers interested in Twitter users’ vaping habits have been surprised by a recent study’s results. Academics in San Diego wanted to analyze data on Twitter concerning the perception and use of electronic cigarettes in the USA. They soon discovered, however, that “pro-vaping” users were almost exclusively robots.
The study, published in the Journal of Health Communication, analyzed 973 tweets randomly selected from a compilation of 194,000 tweets. The tweets selected were published in between October 2015 and February 2016. More than 70% of the tweets analyzed came from bots, which were created to pose as human users. The robots were not shy about singing praises for the electronic cigarette. Indeed, 66% of the tweets analyzed were pro-vaping.
The use of robots on social media platforms is nothing new. Whether fake tweets are used to promote a product or to influence public opinion, the use of bots in many companies is now commonplace. Up until now, however, the use of bots to promote vaping went unnoticed.
Results distorted by bots
Ming-Hsiang Tsou, co-author of the study, believes these fake tweets to be a “real problem” for completing a proper analysis about public opinion concerning vaping. Indeed, it is hard to find out what real people think about e-cigarettes amongst so many fake tweets.
Although some of the tweets’ awkward language and turn of phrase are easy to point out, other tweets aren’t so easy to detect. Certains robots are able to perfectly imitate human language, which has led to many erroneous conclusions.
Twitter is defending itself against these recent discoveries, but it as also deleted tens of millions of suspicious accounts. The social media website has also pointed out that creating fake accounts is completely prohibited.
The tweets in question seem to be based on scientific studies, noting the benefits of electronic cigarettes as a healthy alternative to cigarettes. Nonetheless, as Lourdes Martinez, co-author of the study, points out, the amount of fake tweets raises questions. To what degree can fake tweets influence public opinion, especially when it comes to opinions about vaping and public health?