On Saturday afternoon, my phone dinged with a semipanicked message from a friend.
“I need help from a science boy,” it said. Another ding. “I am seeing reports that‘contains HIV insertions’.” More typing followed.
“A) What the fuck does this mean?” Ding. “B) is it just hysteria?”
The novel coronavirus, 2019-nCoV, public health emergency of international concern. It has induced a sweeping hysteria across social networks like YouTube, TikTok, Twitter and Facebook, largely driven by misinformation and conspiracy theories. Hoaxes leap from user to user like a virus all its own.. At the end of January, the World Health Organization declared it a
My panicky yet mostly reasonable friend was a little confused. Those “reports” largely emanated from Twitter and from one particular user: Eric Fiegl-Ding, a Harvard epidemiologist, who had tweeted about a preprint paper uploaded to the bioRxiv server.
Preprint papers are those that haven’t been reviewed and vetted by other scientists, and servers like bioRxiv act as free and open online repositories for them. When a manuscript is ready, researchers can upload their findings as simply as posting a tweet or dropping a photo on a Facebook feed.
In the last decade, scientists have increasingly looked to websites like bioRxiv to bypass the traditional, laborious process of publishing in journals. This allows them to disseminate their findings much more quickly, collaborate with other researchers and get instant feedback on their work.
Scientists and the media have wrestled with the implications of this new ecosystem for a number of years, particularly in the biological sciences, discussing the pros and cons of the open-access system and the potential for it to confound, overhype or distort scientific findings. Many have championed the preprint system, others have argued against it in favor of traditional publishing methods.
But when 2019-nCoV erupted, it wasn’t the publishing methods scientists had to wrestle with: It was social media.
‘A very intriguing new paper’
On Jan. 31, a research paper uploaded to bioRxiv claimed 2019-nCoV contained similarities to HIV, the virus responsible for AIDS. In short, the paper suggested amino acids found in the novel coronavirus matched those found in HIV. It quickly caught fire online, largely driven by a viral tweet thread constructed by Fiegl-Ding.
“A very intriguing new paper investigating the aforementioned mystery middle segment w/ ‘S’ spike protein: likely origin from HIV,” Fiegl-Ding tweeted, linking to the bioRxiv preprint. A followup tweet began with “WHOA” and then another said “evidence suggest [sic] that 2 different HIV genes are present in the #coronarvirus [sic] S gene region.”
Scary stuff — but not so scary if you just scrolled down the page. Comments quickly piled up on bioRxiv suggesting the paper was flawed and the conclusions were rubbish. Jason Weir, a biological scientist at the University of Toronto, was one of the first researchers to comment, stating very emphatically that the report should be treated with skepticism.
Weir used a publicly available tool, known as BLAST, to disprove the paper’s suggestion that HIV inserts occur within the coronavirus. It took “10 minutes to determine this was not serious science,” he said.
But the study, and Fiegl-Ding’s tweets, generated panic online and helped fan the flames of conspiracy theories suggesting the novel coronavirus was created in a lab, potentially as a bioweapon, and that it was somehow released on an unsuspecting city. The preprint paper was withdrawn by the authors a day later, and Fiegl-Ding deleted his tweets, but the damage was done: Rxivist, which tracks the top trending preprint papers and topics on bioRxiv, shows that the paper is the most tweeted-about preprint of all time.
Fiegl-Ding didn’t respond to requests for comment for this story.
“The response to the HIV/coronavirus paper was unprecedented,” says Rich Abdill, one of the creators of Rxivist. The incredible attention the paper received has hardly slowed down. In response, John Inglis, one of the co-founders of bioRxiv, tweeted on Feb. 1 that the site had added a cautionary note to all uploaded manuscripts.
“bioRxiv is receiving many new papers on coronavirus 2019-nCov,” it begins. “These are preliminary reports that have not been peer-reviewed.
“They should not be regarded as conclusive … or be reported in news media as established information.”
In this case, there were no reports in reputable outlets concluding the paper was correct in finding a link between 2019-nCoV and HIV. Fortunately, the idea didn’t hit the mainstream press because it was completely bogus. Eventually, the preprint was retracted — “a win for science,” as this STAT piece perfectly lays out. The system worked.
Where it didn’t work was social media.
Writing for Nature in July 2018 Tom Sheldon, senior press manager at the UK’s science media center, suggested that preprints could promote “confusion and distortion” and that there are “substantial risks for the broader community.”
“As soon as research is in the public domain, there is nothing to stop a journalist writing about it, and rushing to be the first to do so,” Sheldon wrote.
Many scientists quickly pushed back against Sheldon’s piece, arguing that preprints allow increased scientific collaboration, are no more prone to error than traditional publishing methods and, so long as journalists are responsible in their reporting, wouldn’t distort the public’s view on science. For the most part, that pushback has been right. The preprint process has not been more prone to error or hysteria than traditional publication methods.
In fact, the coronavirus outbreak has demonstrated both sides of science publishing can be just as confusing. On Jan. 30, a paper published in the New England Journal of Medicine proposed that coronavirus carriers can spread the disease even if they don’t show symptoms. But on Feb. 3, members of the German public health ministry wrote to NEJM and showed the report was flawed.
Both the HIV preprint and the NEJM peer-reviewed paper were widely circulated online, particularly on Twitter. The spread validated what Sheldon was saying in 2018, but showed he was pointing his finger in the wrong direction
What Sheldon — and the field of biological science — hadn’t really grappled with at the time was the rise of misinformation on social media. That is the battleground where the public’s view on science is at most risk.
We’ve been battling the spread of misinformation for years. Anti-vaxxers are such a threat that the WHO labeled vaccine hesitancy one of its top 10 health issues in 2019. Flat earthers continue to posit to fervorous YouTube crowds that all of the science we’ve performed about the Earth and the cosmos is wrong. We’ve even had to battle misinformation on social media in previous epidemics, like when Ebola came to the US in 2014 or, just last year, when measles devastated the Pacific Island nation of Samoa.
But we’ve never seen it run as rampant in real time as it hasa been during the 2019-nCoV outbreak.
Where the gold lies
I’ve worked on both sides of the lab bench, as a scientist and as a reporter. But I’d never had an anxious friend ask me about a paper on bioRxiv — until 2019-nCoV erupted.
New science and cutting-edge research can now reach the public immediately. Researchers are able to upload their work to the internet as easily as posting a selfie on Instagram. Anybody can read it, anybody can share it. And just like the best photos or the most replayable TikToks or the most incredible videos, internet stardom is on offer.
As The Atlantic points out, content that exists at the borderline of truth and alarmism is where the gold lies. This holds true for science. It’s this content that has been shared far more than any reasoned, considered take about 2019-nCoV or nuanced scientific explanations of its nature. Nuance is boring. Panic is exciting. It can make social media accounts blow up from 2,000 to 75,000 followers in a few days.
Twitter can turn panic and fear into 15 minutes of fame.
Social media giants are trying to stop the spread of misinformation, but in an outbreak in which so many things remain unknown, it has been more prevalent than ever. We have to sift through mountains of garbage, like a racoon in a dumpster, just to separate truth from fiction. In the echo chamber of social media, the trash has an advantage — it gets amplified and shared and liked and retweeted because it’s a quick way to accrue social capital or spark conversation.
Science and science publishing have been dragged into that world, and they’re struggling to keep up. Science is built on continuous experimenting and testing over months or years. Studies have to be repeated multiple times before they’re accepted as immutable fact. They need to stand up to scrutiny. But we’re now seeing bad studies go viral before scientists even have a chance to read through them.
And so experts will tell us to fact-check stories and ensure our information is coming from verified sources. They’ll say we should only read reputable outlets and be cautious with what we share on social media. They’ll explain how we can spot good science and ignore bad science.
But when the next epidemic rolls around, will that be enough?
Hold on, I’ve just got to check this message.