8 Policy Brief No. 45 Toda Peace Institute has been widely criticised. As with the outside partners on Community Standards, some critics say that Facebook should be doing (and paying for) more of the work internally, not forming partnerships to do it. There are also specific details of the programme’s construction that lead to tensions; all the fact-checking partners are news outlets, which means that a partisan news outlet may find itself fact-checking another news outlet (i.e., a competitor) with opposing political views, as described by Laura Hazard Owen in 2018 in the journalism trade publication Nieman Lab. Overall, it is hard to tell if there is enough scale and speed in these fact-checking efforts to matter, as there is little external information on how well the programme is working. Still, there is more information about the programme in Mike Ananny’s 2018 Columbia Journalism Review article. Labeling and ranking publisher content Facebook has also experimented with labeling publishers who promote their content on the platform, including journalism companies. In 2018, Facebook introduced a Context button, which makes it easy for users to learn more about a publisher when they see a piece of news on Facebook. (At the time, the design team that developed the Context button wrote an admirably transparent Medium post about the process behind the feature.) Facebook has also tested labels for forwarded messages on WhatsApp, a messaging service, as reported by India telecommunications blog Medianama. Specifically, they are trying to use labels to make WhatsApp users more aware that messages forwarded to them are not actually from the individual who forwarded the message. All these labels seem relatively uncontroversial. In contrast, another Facebook initiative to create “trust scores” for publishers caused controversy, because publishers with higher trust scores reportedly get more traffic to their articles than publishers with lower trust scores, and (like everything else about Facebook’s Newsfeed algorithm) the method by which trust scores are calculated is not public. Focus on relationships Partly in response to criticism about spreading fake news, Facebook announced in 2018 that it would decrease the prevalence of articles in the Facebook Newsfeed, and increase content posted by friends. Journalists criticised this move, and it is not clear that anyone else noticed. Viral slowdowns WhatsApp, a messaging product owned by Facebook, was an early testing ground for viral slowdowns — i.e., slowing the speed on viral messages by limiting the number of people that a message can be forwarded to. Previously, a single message could have been forwarded to hundreds of people by one person; now it can be forwarded to only five. In 2019, Jasmine Garsd reported in NPR that the feature, which was tested previously in India, is now going global on WhatsApp. In theory, this could help with problems where people might behave better if they had time to reflect — like slowing down a lynch mob. On the other hand, it could also slow down a message about a genuine disaster — like people at sea warning people on shore about an incoming tsunami.

Select target paragraph3