A new study carried out by the Pew Research Centre has found that the majority of links to content from popular websites are being tweeted by bots, with 66 per cent of tweeted links to popular news and current events websites coming from automated accounts.
In addition, the analysis found that a relatively small number of automated accounts were responsible for a substantial share of the links to popular media outlets on Twitter. The 500 most active bot accounts were responsible for 22 per cent of all tweeted links, compared to just six per cent for the 500 most active human-run accounts.
The study, which focused on data from summer 2017, covered a 47-day period and analysed 1.2m tweets containing links to 2,315 of the most popular websites in terms of shared links to their content.
Overall, bots were responsible for 66 per cent of all tweeted links, but the volume varied depending on the type of content being shared. As one might expect, 90 per cent of adult content came from bots, with sports the next largest at 76 per cent. Commercial product links were 73 per cent bots, news and current events 66 per cent, and celebrity content 62 per cent. Links focused on organisations or groups were 53 per cent bots, and internal linked to other posts on Twitter ranked the best, at 50 per cent automated.
Despite the controversy and concern that has surrounded social media's impact on politics, the study found that among news and current events sites, political content actually ranked the lowest, with 57 per cent of links to sites with a dedicated politics section or political stories among the top headlines coming from bots. Instead, news aggregation sites were the most popular in this section, with 89 per cent of all links to sites of this kind coming from automated accounts.
"We hope that these findings will help illustrate the extent to which bots play a really prominent and pervasive role in posting links to prominent sites and help inform more broadly about automated accounts," said Aaron Smith, associate director at Pew. "It is also worth highlighting that we are in no way implying that the posts that we looked at are illegitimate or problematic because they are automated."