Twitter – the social network with the greatest influence on public opinion despite a comparatively small number of users. If you want to manipulate the public, the short message service seems to be a good starting point. Because it fits together so nicely, there have been regular discussions for more than ten years about how many “real”, i.e. human users, Twitter has – and how large, in contrast, is the proportion of profiles behind which only automated ” Bots” (short for robots). Along with “trolls” (malicious human users), these are believed to be the root cause of all the ugliness of Twitter, from vote-rigging to fake news.
The proportion of bot-controlled profiles in the short message service is not only relevant for Twitter’s credibility, but also for its reliability as an advertising medium: If a significant part of the accounts were not operated by humans, Twitter would have significantly worse chances of advertising at high prices for sale.
There is little that is new in the discussion. But that doesn’t seem to stop billionaire Elon Musk, who has signed an agreement to buy and delist Twitter, from digging it up again. In a letter published on Monday by his lawyers to Twitter’s chief counsel Vijaya Gadde, it said the company refused to provide him with data for his own research on the number of fake accounts behind bots. That’s why he reserves the right to withdraw from the deal. What Musk intends to do with this is unclear. Observers are hesitating whether Musk primarily wants to lower the agreed price for Twitter and renegotiate it, or whether he is looking for a back door to let the deal burst completely.
A discussion that keeps coming back
Twitter leadership has made its position clear: the service itself estimates that fake accounts make up less than 5 percent of the user base. The company speaks of 229 million daily users that the service can reach with its advertising. The fake accounts identified by Twitter have already been deducted. CEO Parag Agrawal also emphasizes that the service blocks more than half a million spam accounts every day – usually before users see them. They will continue to “share data cooperatively with Elon Musk,” Twitter said in response to Musk’s allegation. Estimates of the number of fake accounts from outside the company cannot be made seriously, Agrawal warned.
This is precisely the problem, which keeps the debate coming back. There are estimates and scientific studies on how many Twitter profiles are real and how many aren’t, but you can’t reliably count the fake accounts. Musk himself assumes that up to 20 percent of Twitter accounts are actually fake. A joint study by Indiana University and the University of Southern California in America concluded in 2017 that between 9 and 15 percent of Twitter profiles were operated by bots. That would be significantly more than Twitter admits. But the study results are not really reliable either.
The researchers developed their own program for their study, the “Botometer”. Based on a profile’s followers, their tweeting, retweeting, and following behavior, and the way their messages are phrased, it wants to infer whether the account is run by a bot or a human. The problem: The algorithm classifies many accounts of politicians and organizations that are proven to be operated by humans as bots.
Despite the relevance of the question, Musk is unlikely to be able to find an answer to the question of how many Twitter profiles are real. Whether the takeover goes through is more likely to depend on whether he was serious about his offer – otherwise the dispute could go to court.
