Twitter Bots and Hate Speech in Persian Gulf Countries

The next speaker in this AoIR 2017 session is Mark Owen Jones, whose focus is on social media propaganda in Persian Gulf states. Overall, there is still a considerable lack of research into social media propaganda in Arabic; in Gulf states, there is a long history of 'fake news' in social media, and hate speech towards particular groups, ethnicities, and countries is not uncommon. Hate speech may be operationalised by ruling autocrats as a tool to divide and rule the population; different religious groups are allowed to attack each other, to keep them from uniting and toppling the government.

The present study examined tweets in various topical hashtags in the Gulf since May 2016, and analysed the profile data of the accounts that posted them; this found a number of unusual activity and identity patterns. The research approach was also kept simple in order to make it accessible to civil society and activist groups in the region.

The study found that the spamming of country and regional hashtags with propaganda was common; this undermines these hashtag as spaces for the dissemination of legitimate, useful news. Automated tweets were also used to game trending topics, especially in the context of the recent Qatar crisis. Further, infographics were also used widely to share propaganda content.

One bot network that Mark identified was especially focussed on sharing links to the Saudi 24 satellite network; it pushed xenophobic tweets hostile to Iran, Israel, and Shia Muslims. Bots often accounted for some 50% of all tweets in location-referencing hashtags (such as #bahrain, #iran, #yemen, etc.). Another network pushed the #TrumpWillDestroyIran hashtag, and often presented images of Trump as a crusader knight come to attack Iran. Such bot networks also impacted considerably on the Qatar crisis; anti-Qatar bot activity preceded the crisis itself.

Mark has now created a bot that automatically tweets at accounts that post hate speech, which asks those users whether they feel that such tweets are sectarian. Users often do acknowledge this sectarian nature of these tweets, but also justify them by saying that they are only attacking targets that deserve it, and/or by denying that the Shia users they attack are indeed 'proper' Muslims.