Something is rotten in the state of Twitter: how community notes have spread disinformation about the beating of a transexual woman in Italy

On the 24th of May, a video portraying police brutality against a transgender woman in Italy became viral on Twitter. In the video four members of the city police surround the woman, who sits still on the ground, and hit her repeatedly with kicks and batons. The event occurred in Milan, on the 24th of May itself.

Almost immediately, different versions of the facts started circulating on media and social networks, but on Twitter, on the morning of May 26, community notes appeared claiming that «the transgender individual being subdued by the police had been reported after exposing his penis to school children and threatening to infect people with HIV. He then violently attempted to evade arrest». Twitter community notes are now available for the English language, and they were used in this specific occasion because an Italian journalist tweeted about this story in English.

As it was reconstructed by Facta news, an italian independent fact-checking outlet, Twitter’s community notes’ version follows entirely the one provided by the police union SIULP, initially confirmed – in part – by the Milan city councilor for security, and mention as sources Reduxx (a self-declared feminist news outlet, with many articles attacking Lgbtq+ people), a local newspaper and a Tv program, that reported the police union’s version (the first to circulate publicly in the hours immediately following the event).

The district attorney’s office gave a different reconstruction of the facts though, already on May 24 evening (so, a whole day before the community notes appeared). In the DA’s office version, the police were called because the transgender woman was screaming, causing public nuisance, not because she was exposing her genitals. The principal of a nearby school and various parents confirmed that the woman was not disturbing children. The woman also gave her version of the story – confirming that she was screaming, allegedly against some drunken men that were bothering her, and that she never interacted with children – and on the 29th of May she pressed charges against the policemen.

The dynamic of the facts is still to be verified and most likely it will be necessary to wait for a trial, with hearings and evidence, before having a clear picture of what happened. The question arises on why Twitter community notes were used in that stage of the situation, to convey information not yet fully verified. On the 29th the community notes were removed from Twitter, but only after a very large number of users had already seen them.

Should community notes be Twitter’s main instrument to counter mis- and disinformation, this would be extremely worrying. The methodology is not solid and the outcomes are already mediocre. Applying this tool to the EU languages can potentially result in very negative outcomes, also considering the relatively small community of speakers for some languages (e.g. Bulgarian, Slovak, Hungarian etc.), which will potentially make the scrutiny more difficult by the online community, and the already worrying level of disinformation circulating in some of those media ecosystems.