Big Tech fails once more to contain disinformation in Brazil’s elections

Despite Meta announcing the removal of 600,000 posts during the official campaign ahead of Brazil’s first-round general elections, data analysis consulting firm Novelo Data has identified a string of misleading content published on major social platforms over the last two weeks, which Big Tech firms have failed to curb.

In the three days after the first-round vote on October 2, Novelo Data found 16 live broadcasts and 137 videos posted on YouTube alleging fraud in the electoral process. The posts racked up at least 3.3 million views.

On Facebook, a rehashed 2018 video depicting pro-Bolsonaro lawmaker Carla Zambelli making unfounded electoral fraud claims amassed 791,000 views during the 12 hour period it remained available.

Meta said it removed 310,000 posts violating its policy on inciting violence, and another 290,000 in relation to its terms of use regarding hate speech. It also announced strikes for content violating rules on electoral interference, but provided no concrete numbers to this end.

Experts complain that while oversight has allowed for the removal of potentially harmful content, it remains a case of closing the stable door after the horse has bolted — while these posts remain online, they rack up large numbers of views and interactions before they are eventually pulled.

YouTube has come under special scrutiny for its failure to block recommendations for videos containing misinformation. For instance, a search for “fraud” will bring up suggestions for “fraud in the 2022 elections,” while users searching for “ballot boxes” are recommended to click “rigged ballot boxes.”

The platform claims it has “a solid set of policies and systems that, as well as giving visibility to trustworthy content, reduce the spread of misleading information and allow for political debate.”

• #fakeNews
• #tech