Twitter banned Fátima (@fatimabot), Aos Fatos fact-checker bot, to the platform's API (Application Programming Interface), which consists of a form of communication among computer applications. In practice, the decision prevents Fátima from working to detect the dissemination of links with uninformative content on the platform and from sending users who shared them the fact-checking made by the Aos Fatos team.
In addition to preventing Fátima's operation, last Monday (2), Twitter suspended its profile for more than 10h . The measure was reversed after Aos Fatos contacted the platform. Both the suspension and the ban on the access to the API occurred without any prior notice to Aos Fatos.
In a statement, Twitter reported that "the account was suspended for violating our automation rules, due to the use of an app that mentions or comments in a duplicate way content related to accounts that are not directly interacting with the profile". According to the platform, after analysis, "Twitter withdrew the suspension from the account, but suspended the app for violating its spam rules".
Without the connection to the Twitter API, @fatimabot loses its purpose. Fátima is an awareness bot, that is, it alerts users that the information shared by them contains disinformation. It was programmed to not harm the legitimate debate of the network, and it has mechanisms to avoid spam behavior and to not inflate conversations from malicious bots. Moreover, it does not respond with media, insults or swears, but empathetically, with verified information.
"Fátima is recognized worldwide as an example in the combat of false information through the use of artificial intelligence. Instead of supporting it, Twitter decides to derail it while life-threatening lies multiply on the platform. Along with the non-compliance to its misinformation combat policies, the company facilitates the work of professional fraudsters", says Tai Nalon, Executive Director of Aos Fatos.
Fight against anti-vax speech. The last tweets sent by Fátima to users who shared disinformation refuted the rumor that a Japanese study found mRNA nanoparticles in people vaccinated against Covid-19 with the immunizer manufactured by Pfizer. This disinformation circulated in early June and appeared on Twitter in the last few days. Its reappearance was detected by Fátima, and the users who shared the content were informed that it was false information.
In the last months, Fátima has also alerted users who shared other disinformation about vaccines, such as the false ban on vaccinated people from traveling by air or the claim that vaccines create new variants of Covid-19.
Besides disseminating conspiracy contents about vaccines, one of the Twitter profiles that had their posts answered by Fátima also defended the coup d'état and the end of TSE (Electoral Superior Court), besides posting QAnon theory. The profile is still on the platform, and also appeared in a June report from the Aos Fatos Radar that showed anti-vaccine posts that violated the Twitter rule against disinformation had reached more than 147,000 interactions in three months.
On March 1, Twitter stated that potentially deceptive publication on vaccines would be removed or labeled, and users that shared them would be punished with suspensions and bans. The tweets answered by Fátima are still on, and the fact-checker bot comments are, in most cases, the only warning that that is, in fact, deceptive content.
Good bot. Fátima was created in July 2018 under the premise that all the people exposed to disinformation deserve access to verified information as well. The app was idealized by the journalist and Insper Professor Pedro Burgos and the Executive Director of Aos Fatos, Tai Nalon. The project won the Claudio Weber Abramo Data Journalism Award in 2019.
On Twitter, the profile synchronizes a bank of false or distorted claims already checked by Aos Fatos and automatically maps the social network in search of posts with uninformative links. When finding them, it answers to the profile that shared the deceptive content with a link to the verified information.
