/ 13 June 2019

New app developed to detect Twitter bots — in any language

An estimated 5-10% of Twitter users are bots which account for about 20-25% of tweets posted.
An estimated 5-10% of Twitter users are bots which account for about 20-25% of tweets posted. (Kacper Pempel/Reuters)

South Africa has experienced an upsurge of Twitter bots in recent years, aimed at influencing political discourse and public opinion with seemingly no way to recognise them when they tweet in a language other than English. However, a new tool will allow the detection of these bots, regardless of what language is used.

A Twitter bot is a type of software that controls a Twitter account and can autonomously perform actions such as tweeting, re-tweeting, liking, following, unfollowing, and direct messaging.

A new application, developed by researchers at the University of Eastern Finland and Linnaeus University in Sweden, is able to detect auto-generated tweets independent of the language used relying on machine learning.

Machine learning is a branch of artificial intelligence that uses methods of data analysis to perform a task without using explicit instructions but patterns and inference instead.

According to a statement by University of Eastern Finland professor Mikko Laitinen, an estimated 5-10% of Twitter users are bots which account for about 20-25% of tweets posted.

“The popularity of Twitter as an instrument in public debate has led to a situation in which it has become an ideal target of spammers and automated scripts,” Laitinen said in a statement.

Laitinen notes that bots are relatively harmless but can also be harmful if they are used to spread fake news. The professor says this is why increasingly advanced tools for social media monitoring are critical.

“This is a complex issue and requires interdisciplinary approaches. For instance, we linguists are working together with machine learning specialists. This type of work also calls for determination and investments in research infrastructures that serve as a platform for researchers from different fields to collaborate on,” Laitinen said.

In developing the app, a team of researchers captured 15 000 tweets in Finnish, Swedish and English for analysis. The Finnish and Swedish tweets were mainly used for training of the application, whereas the English tweets were used to evaluate the language independence of the app.

According to the statement, the app is able to process “vast amounts of data quickly and relatively efficiently.”

An example closer to home of the role of Twitter bots can be seen with the controversial Gupta-Bell Pottinger saga where social media was used to manipulate Twitter users.

South African readers will be very familiar with the firm, given its role in pushing a hate-filled propaganda campaign on behalf of the controversial family — a campaign which would ultimately lead to the firm’s collapse.

From January 2016, Oakbay Investments, a South African company owned by the Guptas, hired the PR firm to manage its reputation and corporate communications.

The company — which once bragged about its ability to “drown out negative content” — created a media campaign focused on reducing criticism of the Gupta family and their alleged connection to former president Jacob Zuma. Part of this campaign was to punt the narrative of a ‘white monopoly capital’ where a handful of white business people controlled the most economic means and even the South African media which was critical of the Guptas.

A group of South African journalists, including Ferial Haffajee, Peter Bruce and Adriaan Basson — went on to lodge a defamation claim against AIG Europe, the insurer for now-defunct Bell Pottinger.

READ MORE: SA editors launch defamation claim against Bell Pottinger over ‘WMC’ campaign