Tech Giants Urged to Clamp Down on Misinformation in Spanish and Other Languages

When the Big Three, which is to say, the CEOs of Facebook, Google and Twitter, testify before a U.S. Senate or House committee — as they do again on Thursday — they will face an increasingly familiar litany of challenges from lawmakers.

Top among them: Why do these companies, among the wealthiest and most technologically sophisticated in the world, have such trouble scrubbing their platforms of misinformation, disinformation and hate speech?

Some lawmakers are also asking: Why do these companies have such trouble scrubbing their platforms of objectionable material in languages other than English?

“Think Facebook is failing with disinformation in English? You won’t believe what’s happening in Spanish.” So says a video launched recently as part of a campaign called #YaBastaFacebook, or “Enough Already.”

Mounting Pressure

The campaign is organized by a coalition of advocacy groups, including the Center for American Progress, Free Press, the National Hispanic Media Coalition and the Real Facebook Oversight Board, a British-based group of journalists, filmmakers and others critical of Facebook.

Collectively, they claim Facebook’s efforts to scrub Spanish-language content of political and health-related misinformation and disinformation have been underwhelming.

U.S. Rep. Tony Cárdenas, D-Los Angeles, is on board with the campaign, because, he says, this issue is as personal to him as his 78-year-old mother-in-law.

“My mother-in-law is saying, ‘Is it true that there’s some kind of an electronic thing that they’re going to put in your body?” he said during a press conference the coalition held last week.

Cárdenas was referring to one of the most popular myths that has spread across the internet, in multiple languages: that the COVID-19 vaccine contains a microchip that allows the government to track people. Cárdenas says Facebook — by far the largest social media company, which owns three popular platforms, including WhatsApp and Instagram — has the money to do more, and should do more, before federal lawmakers force it to do more.

“It should not be left to the not-for-profits and to the community at large to do the job Facebook admits that they actually do in English. I'm not saying that they do a good job in English, but the job that they're doing in Spanish and in other languages is almost nonexistent by comparison,” Cárdenas said.

According to a recent study published last year by the human rights nonprofit Avaaz, just 29% of misinformation in Spanish is flagged on Facebook, compared to 70% of comparable material in English.

The study found, for example, that it took 22 days for Facebook to flag a video in Spanish that falsely claimed the virus was deliberately created in a lab at the Institute of Virology in Wuhan. It took nine days for the video to be fact-checked and another 13 days for the company to add the misinformation warning.

By that time, the video had already been viewed 33 million times.

‘Aggressive’ Steps

Facebook spokeswoman Dani Lever told KQED in an email that the social media giant is taking “aggressive” human and automated steps to fight misinformation in Spanish and dozens of other languages, including by removing millions of pieces of misleading content related to COVID-19 and the vaccine.

The company, she added, wants “to continue our dialogue with these [advocacy] groups to strengthen our approach.”

Representatives from YouTube and Twitter said much the same thing.

“Our community guidelines and approach to misinformation apply to all content in any language. And we have reviewers and raters in all relevant languages, looking at e.g. Spanish videos when flagged,” said Elena Hernandez, a spokeswoman for Google, whose parent company owns YouTube, said in an email.

Twitter spokesman Trenton Kennedy wrote that his company's goal was to eventually use both automated and human review to address content that violates its COVID-19 vaccine misinformation rules.

“Machine learning and automated language processing takes time to be effective,” he wrote. “As such, we will begin with English-language content first and use this same process as we work to expand to other languages and cultural contexts over time.”

But as a growing list of incidents around election tampering, racially motivated attacks and COVID-19 vaccine hesitancy clog news headlines worldwide, the heat is on Silicon Valley companies to up their game substantially.

Increasingly, social media giants face litigation, legislation and regulation across the globe. Critics argue that, at best, platforms aren't doing enough to moderate toxic content. At worst, they say, platforms are intentionally amplifying toxic content to profit off of the advertising revenue it brings in, a process they've dubbed the “hate-for-profit” economic model.

That growing backlash comes even as these platforms, which treat much of their data as proprietary intellectual property, make it difficult for journalists, academic researchers and advocacy groups to keep track of problematic content and efforts to control it.

“We are unable to say how significant the harms are, and therefore, whether there is enough being done to combat them,” said Ivan Sigal, executive director of Global Voices, a citizen media network that operates in about 40 languages. “We do not have visibility into the companies.”

Reaching Across Borders

Global Voices conducted an in-depth, international study of popular COVID-19 content in a dozen different languages. One takeaway: The success of a narrative that started in the U.S. is often directly correlated to local political and cultural dynamics in other countries.

“It's hard to isolate or just say that a certain kind of misinformation exists only in a single country and in a single language,” Sigal said.

He offers examples from Bolivia, where one of the most popular coronavirus-related disinformation messages on social media is the scientifically disproven suggestion, proliferated by former President Donald Trump, that people take the drug hydroxychloroquine to protect themselves from the virus.

“It was entangled in a very interesting way, in the political tensions between the traditional right-wing political parties in Bolivia, and the indigenous communities that were actively the source of those kind of natural healing approaches,” Sigal said. “And it all hinged on the mistrust between those two communities. And it was happening both in Spanish, in Aymara and in other indigenous languages.”