Online discourses about (institutional) racism in Rotterdam
— Municipality of Rotterdam
The murder of George Floyd in May 2020 culminated in a global protest movement against racism. On June 5, 2020, Rotterdam was the stage for a large-scale protest. The protesters spoke out against police brutality against the black population in the US while also emphasizing Dutch excesses of racism, including the county’s history with slavery and the appearance of ‘Zwarte Piet’ (Black Pete, the servant of Saint Nicholas). On social media, the protesters received support but faced fierce criticisms too. For the municipality of Rotterdam, New Momentum studied how Twitter users discussed racism throughout 2020.
We focused on excesses of institutional racism in Rotterdam. The goals of this investigation were threefold:
- Identifying online communities speaking out about racism on Twitter while mapping their opinions and perceptions.
- Finding out which communities remained involved when discarding tweets about the BLM protests, which appeared to be a pivotal media event.
- Identifying systematic attempts to influence the Twitter debate by exposing bot and disinformation networks.
We found three communities of Twitter users: Progressives, Conservatives and Media (Rotterdam). Clearly, the progressive community took the lead in putting the issue of racism on the agenda, while the Conservative community responded.
The network of Twitter users discussing racism. The circles are Twitter users, and their sizes express their relative influence. The lines indicate following patterns, and the colors indicate communities of like-minded users. The labels display the community names, the total number of users in each community, and the proportion of users from each community that wrote racism-related tweets. Color intensity indicates whether users shared racism-related tweets or not.
This often led to heated debates, as can be seen in the semantic tag network below:
- The Progressives succesfully managed to leverage the momentum of #BLM to shed light on Dutch manifestations of (anti-)racism (e.g.: ketikoti [abolishment of slavery], ethnic profiling and racism among the Dutch police force).
- The Conservatives responded by pointing out alleged hypocrisy of the protesters (i.e.: ‘whitelivesmatter’, ‘alllivesmatter’, ‘metenmettweematen’ [measuring with two measures]) as well as the alledged double standards of the government in enforcing the covid measures (i.e.: comparing it to the farmers protests [‘boerenprotest’] earlier that year). Furthermore, the Conservatives managed to hijack the KOZP (Kick Out Zwarte Piet) hashtag by accusing KOZP activist Akwasi of black facism and linked the major of Rotterdam (‘Aboutaleb’) to players in international conspiracy theories (‘Soros’, CIA’).
- The Media (Rotterdam) community hardly got involved as Twitter users in that community mostly shared links to news articles.
The prevalence of conspiracy theories in the Conservatives community led us to explore to what extent the Twitter accounts in the network were real people, bots or so-called astroturfers (astroturf is faux grass and, on Twitter, astroturfing is creating the false impression of grassroots support). Using the Botometer API, developed at the University of Indiana, we probed the extent to which the network included fake accounts.
We found bots in all three communities. However, these bots were mostly harmless news aggregators being used by the mainstream media as well as alternative news outlets.
Among the Media (Rotterdam) community, we hardly found evidence for the presence of astroturfers. In contrast, we did find astroturfers among the Conservatives as well as Progressives, although mostly at the extreme and international ends of these communities.
Bot network: darker shades of red signify accounts that are more likely to be bots. In all three communities we found news aggregator accounts, automatically sharing news articles posted on websites of mainstream as well as alternative media.
Interactions in the bot network. The lines signify retweets, replies, quotes and/or mentions, where thicker lines show repeated interactions. In this manner, thick chains unveil bot networks.
Although we did find evidence of bots and astroturfers in the network, we did not find any cues that they were activated on a large scale to destabalize the debate about racism in Rotterdam. Rather, the debate about racism in Rotterdam seems to take place in the context of an international debate that is influenced by bots and astroturfers among the Conservatives as well as Progressives.