The final speaker at the workshop of the Bots Building Bridges project for today is Florian Muhle, who begins by highlighting the transformation of social media bot detection approaches to take into account a much more complicated and hybrid environment.
Bot detection was already very difficult, and is no universal solution: human users also engage in inauthentic content amplification, for various commercial, political, and other reasons. It is therefore more useful to focus on the effects of such artificial amplification: and here, a continuing focus on single platforms is no longer useful since many such amplification efforts aim at dispersing messages as widely as possible, beyond the boundaries of any one single social media platform.
Like social media overall, then, artificial amplification efforts operate in a hybrid media system: their effects are observed and discussed outside social media as well, and thematised by societal actors including politicians and journalists; indeed, artificial strategies may well target these actors as further message multipliers. Artificial amplification should therefore be studied in the context of multimedia attention dynamics.
One case of this is the case of conservative candidate for German Chancellor in the 2021 federal election campaign, Armin Laschet, who was seen laughing in the background of a solemn press conference about a severe, deadly natural disaster. This was visible in the live news coverage of several TV stations, but not immediately remarked upon; however, after several hours the story emerged to greater prominence, following considerable debate of the event under the #LaschetLacht hashtag on Twitter. This is also because the hashtag became a story in its own right in the German news media.
Twitter timelines show a rapid rise in the use of this hashtag over the hours following the original press conference. Florian’s project explored whether highly active Twitter users played an important role in increasing its prominence, and found a typical long-tail distributional: while most accounts engaged only briefly, a small number of accounts were highly active in discussing the event and promoting the hashtag.
Some such accounts also engaged in ‘hashtag bombing’, deliberately bringing the hashtag to the attention of key journalistic, media, and other public actors. They also contributed to dissemination by retweeting other accounts’ messages, and were in fact so successful in this effort that the hashtag eventually even attracted unrelated spam accounts.
Highly active accounts played a central role in the diffusion of this trend, therefore; they were also very effective in activating their neighbours in engaging in disseminating the hashtag. The hashtag became the top trending topic open Twitter on the day of the incident.
This, in turn, also resulted in considerable media coverage; by the evening, coverage of Laschet’s inappropriate laughter had come to overtake coverage of the overall press conference itself in volume. Laschet’s own response to the Twitter storm, apologising for his faux pas, only generated further coverage, of course.
Notably, party conservative voting intentions declined considerably in the aftermath of the event, and Laschet’s CDU lost the election. Subsequent media reporting pointed to the event as a pivotal moment in the election campaign – such single-issue explanations may well be overstated, however.
