By Ben Kerrigan-
YouTube and Facebook have been allowing disinformation to be spread about Brazil’s election campaign, adding to the bitterness in an already polarised and violent election, according to a new report by the human rights organisation Global Witness.
The NGO produced a series of purposely misleading ads during an election season that has been dominated by the bitter race between far-right incumbent Jair Bolsonaro and his leftist challenger, former president Luiz Inácio Lula da Silva.
AD: Capeesh Restaurant
Some of the mock ads urged people not to vote; others, like Bolsonaro himself, questioned the credibility of the election; and a few gave a false date for the ballots.
YouTube approved all of them to run and Facebook approved half, Global Witness said, although it stressed it withdrew the ads before they were published to avoid spreading confusion.
TikTok similarly approved 90 percent of ads featuring misleading and false election disinformation, which were tested by Global Witness and the Cybersecurity for Democracy (C4D) Team at NYU, the organizations today reveal.
AD: Oysterian Sea Food Restaurant And Bar
Facebook approved a significant number of similarly inaccurate and false ads, while YouTube detected and rejected every single such ad submitted. YouTube also suspended the channel used to post the test ads.
The experiment, to determine how well social media platforms are living up to their promises to stop disinformation that can destabilise democratic processes, posted 20 ads to the three platforms in both English and Spanish language, targeted to “battleground” states like Arizona, Colorado, and Georgia.
Global Witness carried out during the tense Brazilian elections. In August 100 percent of ads submitted for approval to Facebook were approved for publication, that included incorrect information such as the date of voting, as well as denying the credibility of the election and urging people not to vote.
Last month, half of the same ads submitted to Facebook were again approved for publication and in this latest test carried out following the first round of voting, 50 percent of the ads were again approved, including some which had previously been rejected.
The experiment was this time widened to include YouTube, and every single ad tested was approved by the Google-owned platform. Ads claiming that in some locations the run-off date had moved from the 30th to 31st October and that the result of the election was already pre-decided were amongst those YouTube cleared to be published.
“It’s frankly shocking that these massive firms with the technological prowess they clearly have, are unable to weed out such blatant disinformation being pushed onto their users,” said Jon Lloyd, Global Witness’s senior adviser.
“If it wasn’t already obvious it should now be undeniable to even the biggest sceptic – social media firms are fundamentally failing in their responsibility to stop democratic processes being undermined by false, misleading and purposeful deceit.”
The group said Facebook approved all of its 11 ads in late July and early August, two months before ballots for president, Congress and the governors and legislatures of 27 states.
Half of the ads were approved in a second test a month before the 2 October ballots and the same amount were approved after Lula won the first-round election with 48.4% to Bolsonaro’s 43.2%. Some of the ads that passed Facebook’s tests had been rejected in earlier tests, Global Witness said.
That last wave, carried out between Oct. 11 and 13, was expanded to include YouTube “and every single ad tested was approved by the Google-owned platform,” including some giving false information claiming the run-off ballot was being put back a day from Oct. 30 to Oct. 31.
A Meta spokesperson said it invests “significant resources” to protect elections and rejected 135,000 Brazil-related ad submissions between Aug. 16 and Sept. 30.
“These reports were based on a very small sample of ads,” it said of the Global Witness submissions, “and are not representative given the number of political ads we review daily across the world.”
YouTube said it “reviewed the ads in question and removed those that violated our policies,” although the Global Witness report showed all the ads submitted were approved by the Google owned site.
This year’s election is one of the most bitter in Brazilian history, pitting the far-right incumbent against his nemesis who governed for two terms between 2003 and 2010.
It has, like much of Bolsonaro’s four-year term, been pockmarked with nastiness and lies, and both sides are accused of promoting fake news.
The far-right has been particularly guilty, and experts with Netlab, a Rio centre for internet and social studies, said misinformation campaigns are more complex than ever this year, with multiple actors and bots blasting out coordinated messages of lies across different platforms at the same time.
“There is an ecosystem of fake news that is shared in seconds,” said Netlab’s director Marie Santini. “When you get the same message from several different sources at the same time you feel that it’s legitimate.”
This week Brazil’s electoral court told YouTube, Twitter and Google to take down dozens of sites or posts supporting Bolsonaro and said the far-right’s misinformation campaign was “harmful to Lula” and carried “significant repercussions and persistent effects.”
Experts said Brazil was not alone in facing such online challenges to democracy. Across the world social media empires have let bad actors influence ballots through bots, fake news farms and lax oversight and it is a problem that is worse outside the English-speaking world.
“You have to invest not just in AI technology able to understand Brazilian Portuguese but also in qualified Brazilian moderators to filter disinformation,” Santini said. “And they don’t want to do that because it is expensive.”
“But it’s also neglect. They don’t think it matters because it is Brazil. Facebook and Google’s shareholders are in the US, not here.”
AD: Heritage And Restaurant Lounge Bar