Misinformation and disinformation on Facebook has become ever more important as the Covid-19 pandemic ploughs on around the world. The pandemic is leading platforms to enhance their policies, processes and technologies for fact-checking Covid misinformation.
Health misinformation broke onto the tech radar with the US measles outbreak in 2018. There were 371 measles cases in the US that year and a further 1,215 cases in 2019. That was the largest number of cases since the disease was considered eliminated from the US in 2000, thanks to mass vaccination.
How Does Facebook Fact-Check Covid Misinformation?
In the 12 months to May 2021, Facebook removed 16 million pieces of content and added warnings to over 160 million more. YouTube has operated similarly, blocking almost a million videos on the grounds of “dangerous or misleading Covid-19 medical information.”
It’s well-known that such a quantity of information is too great to manually review, and critics often denigrate “bots” or “the algorithm” when they disagree with their content being flagged. Facebook develops its own AI that reviews content by connecting it to already known and identified misinformation.
It’s less appreciated that those original claims are fact-checked before being identified as misinformation. To do so, Facebook and Google both rely on the International Fact-Checkers Network, which includes SciCheck, Metafact, and Science Feedback for scientific review.
One prominent incident was a Wall Street Journal article from February 2021 titled, “We’ll have herd immunity by April.” Though the piece was authored by a US professor of health policy, Facebook tagged it as “misleading.”
SciCheck contacted three scientists who they identified as having published in fields relevant to the article’s claims. Based on their responses, a tag of “misleading” was affixed to the article when shared on Facebook.
Though the editor of British Medical Journal, for one, expressed concern about legitimate scientific debate being considered misinformation, Facebook, the fact-checkers and their helpful pro bono peer reviewers turned out to be completely right: current case numbers in the US are on par with last November.
Some scientists see all of this as tech companies intervening in legitimate scientific debate. In my opinion, that position misreads what Facebook is all about.
If the general scientific consensus is that a finding is misleading or inaccurate, then that is a perfectly fair label for Facebook to give it. Scientists have plenty of fora to debate the methodological and interpretive specifics amongst themselves, where they can do so without giving easily shareable ammunition to opportunists and other misguided people.
Follow Christian on Twitter for more news updates.
Article image courtesy of @priscilladupreez and feature image by @michalmatlon via Unsplash.