Lapses in tech companies’ policies to address Spanish content led to a proliferation of misinformation targeting Latino voters around Election Day, according to several advocacy groups.
Spanish misinformation campaigns largely mimicked those in English that cast doubt on the security of mail-in ballots, later calling into question the election results.
But while the English-language posts were regularly removed, Spanish ones often “slipped through the cracks,” said Jessica González, co-CEO of Free Press and co-founder of the civil rights coalition Change the Terms.
“I think there were massive failures across the board,” González told The Hill.
González said Facebook in particular failed to address English and Spanish content equally.
“They certainly need teams of people who are native speakers, who understand dialects, idioms, understand casual speech in all the major languages that are spoken and used on Facebook,” she said.
Even when posts are labeled, González said the language used by Facebook “is very weak.”
A Facebook spokesperson defended the company’s actions toward limiting the spread of misinformation in Spanish.
“Ahead of the election, we took a number of steps to combat misinformation in Spanish: we built a Spanish version of our Voting Information Center where people can find accurate information about the election and added two new US fact-checking partners who review content in Spanish on Facebook and Instagram,” the spokesperson said in a statement.
Facebook also limited forwards on its messaging service and launched a feature on WhatsApp, a messaging system owned by Facebook, where users could search the web to receive more information about their chats, the spokesperson added.
Facebook CEO Mark Zuckerberg was asked during a Senate Judiciary Committee hearing Tuesday what steps his company will take to improve content moderation for Spanish-speaking communities before runoff elections in Georgia on Jan. 5 that will decide which party controls the Senate next year.
“This is something that we are already working on and worked on ahead of the general election,” Zuckerberg said in response to the question from Sen. Richard Blumenthal (D-Conn.).
“We’re certainly committed to focusing on this,” he added.
A Twitter spokesperson said the company has dedicated teams of specialists who provide “24/7 global coverage in multiple languages, including Spanish,” as part of its enforcement of its Civic Integrity Policy and guidance on labeling premature election results.
“We have and will continue to enforce our rules impartially to protect the integrity of the conversation around this election,” the spokesperson added.
Critics say they have data showing a lack of scrutiny surrounding Spanish-language posts.
Preliminary data from the global civic organization Avaaz suggest unsubstantiated claims in Spanish were not labeled by Facebook at the same rate as English ones.
Members of President Trump’s legal team, including Sidney Powell and former New York City Mayor Rudy Giuliani, made a series of claims about electoral fraud not supported by evidence, including spreading debunked claims about electoral software manufacturer Dominion Voting’s systems.
Avaaz analyzed 10 of the highest-performing posts on Facebook promoting such claims in both English and Spanish and found that while 5 out of 10 of the highest performing posts repeating the claims from Powell and Giuliani in English had been labeled by Facebook, just 1 out of 10 of the highest-performing posts repeating the claims in Spanish had been labeled.
“False claims of voter fraud are helping plunge the country further into chaos and confusion, and often the most vulnerable communities in the country are paying the highest price,” Avaaz Campaign Director Fadi Quran said in a statement. “Facebook is failing to protect these communities by failing to enforce its anti-misinformation policies equally across all languages.”
Other advocacy groups say Trump’s unwillingness to concede the race to President-elect Joe Biden has led to more posts spreading misinformation.
“We’re seeing a lot of posts in Spanish around the same kind of messaging and framing of the hashtag ‘stopthe-steal’ message,” said Lizette Escobedo, national director of civic engagement at NALEO.
“Stop the steal” has become a rallying cry for supporters pushing pro-Trump election misinformation about voter fraud.
The continuation of election misinformation has pushed NALEO to extend its staff monitoring of social media to search and flag posts spreading misinformation. Escobedo said the organization had initially planned to keep monitoring a few days after Election Day due to expected delayed results.
“As we started observing this messaging come in and started hearing President Trump kind of drumbeat this message we thought, ‘Uh-oh, this is going to come back,’ ” Escobedo said.
Even as NALEO’s team searched for content to flag, though, Escobedo said they ran into challenges to searching for content in Spanish through the social media giants’ tools.
“A lot of these tools are created to kind of work for Spanish, but mostly work for English,” she said.
Misinformation about the election poses some unique challenges for groups aiming to educate voters of color against false or misleading claims.
For example, misinformation targeting communities of color often will use race as a wedge issue, pitting communities against one another to create “different types of racial hierarchy and race supremacy,” said Meghna Mahadevan, chief disinformation defense strategist at United We Dream Action.
Another major issue is impersonation accounts, where users pose as members of a different racial group, she said.
“There’s this whole coopting of identities by someone who is not advocating for the interest of that race,” Mahadevan, who previously worked for Facebook, added.
With advocacy groups pushing Facebook and Twitter to take more action regarding misinformation in different languages, Escobedo said she thought Facebook and Twitter were the “most willing to have communication” with organizations about those issues.
Platforms with less or no content moderation pose bigger challenges, she said, citing Parler, a Twitter alternative gaining traction with conservatives.
Sites that don’t label or fact-check posts could become a space for conspiracy theories to flourish, she said, regardless of the language.