Google Play icon

Online Bots have Significantly Evolved in just 2 Years (between 2016 and 2018), New Study Shows

Share
Posted September 9, 2019

According to a new study published by a team of Italian and U.S. researchers (from the Scuola Universitaria Profesionale della Svizzera Italiana, and the Information Sciences Institute of the University of Southern California, respectively), online bots – or fake accounts enabled by AI – have successfully evolved to better mimic human users within the span of just 2 years.

Writing in the academic journal First Monday, study authors Emillio Ferrara, Luca Luceri, Ashok Deb, and Silvia Giordano provide a detailed account of their efforts to determine the changes, or lack thereof, in bot activities during the U.S. elections of 2016 and 2018.

Securing undistorted communication taking place on social media (as far as such is possible) – and especially reducing the potential damage caused to public discourse by sophisticated online bots – might be one of the key factors in preventing biased election results. Image credit: Today Testing via Wikimedia

Securing undistorted communication taking place on social media (as far as such is possible) – and especially reducing the potential damage caused to public discourse by sophisticated online bots – might be one of the key factors in preventing biased election results. Image credit: Today Testing via Wikimedia.org, BY-SA 4.0

Having examined the online activities of almost 250,000 social media users (30,000 of whom were revealed to be bots) during the aforesaid period, the researchers found that bots have managed to adjust their conduct to match the changing online behaviours of real people.

Whereas in 2016 bots were typically deployed for retweets and posting high volumes of new messages, in 2018 – taking a cue from their flesh and bone counterparts  – bots were found to ditch the aforesaid pattern in favour of a multi-bot approach which is more reminiscent of natural human behaviour as it pertains to a specific idea.

Furthermore, as human users of Twitter shifted their online engagement to replies to other users, bots started developing and posting theme-based polls, presumably to bolster the respectability of the associated accounts (in many people’s eyes, polls are a sign of respectability, as they are often used by trusted news agencies and pollsters).

“Our study further corroborates this idea that there is an arms race between bots and detection algorithms. As social media companies put more efforts to mitigate abuse and stifle automated accounts, bots evolve to mimic human strategies. Advancements in AI enable bots producing more human-like content,” said lead author Emillio Ferrara.

According to Ferrara, understanding how bots evolve and developing strategies for detecting the more sophisticated kinds thereof is of paramount importance to ensure a bona fide democratic process during the upcoming 2020 US elections.

Source: eurekalert.org

Featured news from related categories:

Technology Org App
Google Play icon
84,049 science & technology articles

Most Popular Articles

  1. Efficiency of solar panels could be improved without changing them at all (September 2, 2019)
  2. Diesel is saved? Volkswagen found a way to reduce NOx emissions by 80% (September 3, 2019)
  3. The famous old Titanic is disappearing into time - a new expedition observed the corrosion (September 2, 2019)
  4. The Time Is Now for Precision Patient Monitoring (July 3, 2019)
  5. Europe and US are Going to Try and Deflect an Asteroid (September 6, 2019)

Follow us

Facebook   Twitter   Pinterest   Tumblr   RSS   Newsletter via Email