Skip to navigationSkip to content

Bots post most of the links on Twitter (but that’s not all bad)

Youtube/Zhejiang TV official channel
More links!
Published Last updated This article is more than 2 years old.

Twitter bots, it seems, are better at their jobs than humans.

The Pew Research Center examined roughly 1.2 million English-language tweets over a six-week period last summer, and found that 66% of links to the internet’s 2,315 most popular websites came from suspected automated accounts. For pornography websites (which Pew calls “adult content”), 90% of links posted on Twitter came from bots, while the source of links to news sites was the exact overall average split between human and bot referrers.

When it came to news, the top bots far outpaced the most prolific human link-sharers. The 500 most active suspected bot accounts accounted for 22% of all shared links for news and current events relatively to just 6% by the 500 most-active human users.

This is not necessarily a bad thing. Though bots may be thought of by many as the work of social media misanthropes, or extensions of the Russian government, many are benign, according to Pew, including automated accounts for news organizations or businesses. For example:

  • Netflix Bot (@netflix_bot): automatically tweets new movie arrivals
  • Museum Bot (@museumbot): random images from the Metropolitan Museum of Art collection
  • CNN Breaking News Bot (@attention_cnn): unofficial CNN breaking news account

There was no “political bias” among automated accounts for the top sites, according to Pew. Bots shared links to political sites popular with conservatives (44%) and liberals (41%) at roughly similar rates.

Pew’s report, however, doesn’t capture the entire picture of how partisan news is disseminated on Twitter. Pew’s study only examined major media outlets, and did not investigate how humans share politically charged news. Other studies over the past two years have detected a strong political valence. A 2018 study by the University of Oxford’s computational-propaganda project, for example, studied about 13,500 politically active Twitter users and 48,000 public Facebook pages. The researchers sought to detect how “junk news” (defined as extremist, sensationalist, conspiratorial, masked commentary, or inaccurate content) was shared prior to Donald Trump’s first State of the Union address last January. In this study, a wider range of “junk news” was shared more often by groups of Trump supporters “than all the other groups put together.”

Correction: This story has been corrected to reflect that Pew updated their report to note that suspected bots shared roughly 41% of links to political sites shared primarily by liberals and 44% of links to political sites shared primarily by conservatives (a difference that is not statistically significant). A previous version of the report switched these numbers between liberals and conservatives. 

📬 Kick off each morning with coffee and the Daily Brief (BYO coffee).

By providing your email, you agree to the Quartz Privacy Policy.