Some falsely claim President Biden caused the baby formula shortage because he gave it all to undocumented immigrants. Others show doctored images of dogs urinating on Trump campaign posters or wrongly claim mail-in ballots have significant voter security issues.
Many immigrant communities trust the apps for information, which can allow lies to take hold and shape opinions among growing voting blocs in states such as Pennsylvania, Texas and Colorado, experts said.
The messaging apps are a key battleground. Meta-owned WhatsApp is popular among Spanish and South Asian language speakers in America. Tencent owned WeChat is popular with millions of Chinese Americans in the United States.
But combating misinformation on these portals can be challenging. Social media giants have pledged to combat misinformation, labeling claims as false or taking posts down altogether. But studies show their technology does worse at moderating content in non-English languages.
Moreover, since WhatsApp is encrypted it limits the amount of data its owner can track, hindering efforts to label messages as misinformation. Political fact checking projects on WeChat can often disrupted, Asian American organizations said.
Immigrant and media groups are trying to do it themselves, using machine learning, chatbots and traditional forms of fact-checking journalism. But the scale of lies are too much, they said, leaving them overmatched.
“It’s possible to debunk the most dangerous lies,” said Jinxia Niu, who runs a Mandarin language fact checking initiative called PiYaoBa in California. “But it’s definitely impossible to combat all of them.”
Representatives from Meta and WeChat did not return request for comment.
Meedan, a nonprofit that builds software for newsrooms, has partnered with four media outlets, including Telemundo and Univision, to combat Spanish-language misinformation ahead of the midterms.
Their software connects a newsroom’s WhatsApp account to a central database. The news organizations publicize a number which readers can call, text or send images or audio messages to, seeing if an article they’ve received is real or not, Meedan chief executive Ed Bice said.
That feeds into an algorithm which analyzes the messages. If the news outlet already has an article debunking that claim, the software will send it. If not, the database will log the potential lie. From there, reporters can access the database to see which lies are taking hold and which might be worth debunking.
Ronny Rojas, an investigative journalist who leads Telemundo’s fact-checking initiative with Meedan, said that in the run-up to the midterm election, he has received 20 to 30 Spanish language messages a day asserting potential lies.
In places like such as Colorado, messages falsely claim Democrats are castrating children and conducting “gender experiments” on them. In Pennsylvania, false messages are circulating about voter fraud regarding mail-in ballots.
Rojas said there are so many lies that debunking them all seems unlikely. But it’s important to try, he said, so people don’t “decide who to vote for based on lies or misinformation.”
Neil Makhija, who heads the advocacy group Indian American Impact, said his idea to combat misinformation targeting South Asians on WhatsApp started during the pandemic. He noticed messages circulating among family members that covid-19 could be cured by ingesting turmeric.
His organization created Desifacts.org, an initiative that dispels political lies spread on WhatsApp. The group created a WhatsApp tip line, where South Asians can share content going viral and get it fact-checked. True articles are rated as “Honest-Tea,” while completely false stories are labeled “Masala chai.”
The organization’s staff are in more than 200 WhatsApp groups, monitoring the spread of disinformation and fact-checking in real time, if needed. The group is creating an artificial intelligence fueled chatbot, so people who submit articles, memes or messages for a fact check can get an immediate response with an article or graphic his team has already created.
Niu, of the civil rights group Chinese for Affirmative Action, said her service, PiYaoBa — or “let’s fact-check it” in Mandarin — is one of the few organizations focused on dispelling election lies in Mandarin on WeChat.
Launched this year, the group monitors WeChat — which has roughly 3.8 million monthly active users in the United States — for misinformation that might be spreading among Chinese Americans. Many of the falsehoods, Niu said, circulate among people in California and New York.
Two false claims that have been by identified by PiYaoBa include: the “Deep State” was why police didn’t respond quickly to the school shooting in Uvalde, Texas, and the FBI raid on Donald Trump’s home in Mar-a-Lago was orchestrated by the Democratic Party to rig the midterm elections.
Niu’s team of three full-time and 15 part-time staff are trying to debunk the most popular lies in real time, publishing articles and posting them on WeChat channels.
They have written roughly 100 fact-checking pieces and written 91 disinformation alerts. But Niu said that’s not enough, since around 138 accounts on WeChat are dispelling political misinformation and propaganda.
“We’re definitely under-matched,” she said.
