Over the past week, Twitter has flagged dozens of tweets with factual information about covid-19 as misinformation and in some cases has suspended the accounts of doctors, scientists, and patient advocates in response to their posts warning people about the illness’s dangers.
In interviews with The Post, Twitter acknowledged the problem. The company removed the labels and restored the accounts after queries about 10 specific tweets and accounts.
“We’re always working to improve the safety of our service and ensuring we provide avenues of recourse when we get it wrong through our appeals processes,” Twitter spokesperson Celeste Carswell said. “We acknowledge the mistakes made in these cases, and we are reviewing our team’s protocol to safeguard against such mistakes in the future. We appreciate the community’s feedback and remain focused on reducing harm and providing informative context across Twitter.”
Some users received messages from Twitter apologizing for the mistake. “Our support team has reviewed your account and it appears we made an error,” one email reviewed by The Post said. “We’ve determined that the Tweet(s) is not in violation of our Covid-19 Misleading Information Policy that therefore your account should be restored to full functionality. We sincerely apologize for any inconvenience and hope to see you back on Twitter soon.”
Johanna Po, an oral pathologist with a PhD in molecular biology who had one of her tweets flagged last week for misinformation (the flag was later removed), said discussion of the misapplied flags has been a big topic in the science and medical communities on Twitter. “It’s happened to quite a number of medical people and people in bio research,” she said.
It was unknown how many tweets or accounts have been affected by the mislabeling, but Po said that even if Twitter’s actions affected relatively few account holders, the wrongly applied flags undermine the authority of the science and medicine community and weaken efforts to counter vaccine or covid misinformation.
“In terms of public perception, it puts into question our credibility as scientists," she said. “If we put something on Twitter like a research study, and you say it’s wrong, it invalidates everything we’re saying.”
Emily Vraga, an associate professor at the University of Minnesota who specializes in misinformation, said Twitter’s recent errors confuse people and feed conspiracy theories.
“We know that when accurate information is labeled as false, people are less likely to think it’s accurate, even when it’s shown to be true,” she said. She added that long-term misapplication of the label could also lead users to distrust Twitter’s misinformation systems entirely, rendering the label useless.
Throughout the pandemic, Twitter has struggled to contain misinformation on its platform. In April 2020, Oxford University researchers released a study showing that nearly 60 percent of false claims about covid remained online without a warning label. Anti-vaccine sentiment has also thrived on Twitter. A February report by the think tank the German Marshall Fund found that vaccine skeptics significantly outperformed news outlets on Twitter. “Anti-vaccine promoters managed to evade platform policies, distorting the information system with out-sized footprints during the second half of 2021,” researchers wrote.
When precisely the current problem began is uncertain, but it has been a subject of conversation among medical professionals and scientists on the platform since at least last Thursday, with tweets calling attention to the issue garnering thousands of likes and retweets.
Some reports suggest the issue began before then.
On Aug. 16, Henry I. Miller, a physician and molecular biologist who worked for 15 years as a medical reviewer at the federal Food and Drug Administration and as a research associate at the National Institutes of Health, wrote about his experience after he posted a headline about new vaccine research. The tweet, which he said he’d posted July 28, was labeled misinformation.
Miller deleted his tweet to re-access his account, but he said the platform should apologize to him and other pro-vaccine, pro-science communicators who have had their tweets flagged. “It shows you how disorganized the monitoring of disinformation on Twitter is," he said. “They’re just in disarray.”
Another user, Anita Cheng, a former molecular biologist who was embedded in San Francisco’s emergency covid response team, said one of her tweets promoting vaccination was mistakenly flagged two weeks ago.
And Chantal Moore, a science and technology writer in Vancouver, said several tweets of hers sharing information on how to keep safe by wearing N95 masks had been flagged as far back as June. Moore didn’t file appeals on her previous tweets, but the flag on one June tweet, which cited information from a CDC Instagram post, was removed after The Post brought it to Twitter’s attention.
Other Twitter users were temporarily suspended or had their tweets flagged after sharing CDC data on childhood covid deaths, posting headlines and links to studies in highly regarded scientific journals, warning of the dangers of covid to children, and sharing a quote from a New Yorker story that highlighted covid’s danger to pregnant women.
Several tweets that were flagged as misinformation amplified studies related to long covid, a devastating condition that affects millions and for which there is no cure. Biden announced last July that long covid is a federally recognized disability, though many have struggled to obtain disability benefits.
Farid Jalali, a gastroenterologist in Southern California, had a tweet flagged that was reacting to an article in the scientific journal Nature about the continued risk of cardiovascular problems even many months after a covid infection resolves. The label was later removed, but he said such errors make it harder for physicians to keep their patients informed.
“The evidence has become clear that recurrent covid-19 infections can be harmful to the health of an individual, and therefore it is part of physicians’ job requirement to ensure the public is aware of those risks,” he said. He noted that many doctors miss covid’s connection to strokes, heart attacks, and blood clots because they think of the virus as solely attacking the respiratory system, when, in fact, covid affects many organs throughout the body.
Helga Gutmane, cofounder of the Long Covid Research Initiative, a nonprofit group that promotes research into long covid, said wrongly applying misinformation labels to tweets makes it harder to get scientific information about long covid to the public. On Saturday, Twitter flagged as misinformation a tweet the initiative posted linking to a paper in Nature on covid transmission from mothers to babies. Twitter later removed the label, but the mislabeling confused her followers, Gutmane said.
Apparently, @TwitterSupport is freezing people's accounts for tweeting research studies, now.— It'sME(Jaime) (@exceedhergrasp1) August 18, 2022
Hey, @Medscape, were you aware that your scholarly article is being labeled "COVID disinformation" and that sharing it is apparently a bannable offense? pic.twitter.com/rIfM5vsoow
Doctors said shifting views on covid have led to an uptick in online attacks and perhaps an increase in users flagging content to Twitter as misinformation. “Physicians who are advocating for a more cautious approach to removing covid mitigations, tend to be the physicians who are challenged and reported [for misinformation] and get harassed,” said Graham Walker, an emergency physician whose tweet attempting to fight anti-vaccine misinformation was labeled as misinformation by Twitter last week. The label was later removed.
“It’s an all-around assault on just getting factual information out,” Jalali said. Whereas in 2020 the public seemed united in efforts to conquer the virus, doctors said that now, people across the political spectrum often direct vitriol toward anyone attempting to raise awareness about the often devastating effects of covid-19.
“The physicians that seem to be most challenged and criticized online are the ones trying to represent the people who are still at a pretty high risk for covid,” Walker said.
A report published in JAMA Internal Medicine last year found that one in four doctors had been harassed and attacked on social media. In March, a survey by Science found that doctors and scientists who continued to warn about covid’s dangers received an onslaught of vicious attacks, including death and rape threats. Last month, the American Medical Association published a piece on how harassment of doctors is on the rise, especially online.
“You have 99 percent of the world who don’t want to pay attention to covid and long covid, they want to pretend like it doesn’t exist,” Gutmane said. “It feels like we’re this one small little cog in the machine, trying to stop the Titanic from sinking.”