Coronavirus misinformation slipping through Facebook's ad review system

Ad encouraging users to drink bleach was among those approved to run in new investigation that exposes flaws in automated system.

Facebook: these ads were approved by ad review system
Facebook: these ads were approved by ad review system

Facebook is approving ads to run that contain deliberate and dangerous misinformation relating to Covid-19, according to an investigation by US non-profit organisation Consumer Reports.

The organisation created seven paid ads containing varying degrees of coronavirus-related misinformation to test whether Facebook's system would flag them prior to running – but all got approved.

In order to do this, Consumer Reports set up a fake Facebook account and a page for a made-up organisation called the "Self Preservation Society", through which it created the ads. The work all featured content that Facebook has banned over the past few months, including "claims that are designed to discourage treatment or taking appropriate precautions" and "false cures".

The ads ranged from subtle to blatant misinformation. But even the most flagrant ones – containing claims such as "Coronavirus is a HOAX", that social distancing "doesn’t make any difference AT ALL" and that people can "stay healthy with SMALL daily doses of bleach" – were approved. Consumer Reports pulled the ads before they were able to run.

Facebook confirmed that all seven ads created by Consumer Reports violated its policies.

In a statement, a Facebook spokesperson said: "While we’ve removed millions of ads and commerce listings for violating our policies related to Covid-19, we’re always working to improve our enforcement systems to prevent harmful misinformation related to this emergency from spreading on our services."

Facebook's system can review ads before, during or after they complete their set. It relies primarily on automated review; human reviewers are used to improve and train its automated systems and, in some cases, review specific ads.

It is likely that the seven ads created would eventually have been found and removed. But they could have already caused damage among users by this time. This is particularly concerning when considering Facebook's reach; the parent company operates four of the top six social networks in the world.

With its vast reach, Facebook has been under a lot of pressure over the past few months to stem the spread of misinformation related to Covid-19 on its platforms. But it is attempting to do this with fewer staff, after sending all of its content reviewers home for their safety.

Facebook did flag last month that, due to a reduced and remote workforce, and therefore a heavier reliance on automated technology, this may lead the platform to "make mistakes", specifically resulting in an increase in ads being "incorrectly disapproved". It did not flag the possibility of the opposite of this – ads being incorrectly approved – and the much more dangerous conotations of this.

A version of this story first appeared on Campaign Asia-Pacific