Facebook fights against fake news, but can it beat internet’s ‘bad guys’?

A couple of years ago, Greg Marra, product management director at Facebook, spent most of his time figuring out how to make the site’s News Feed more enticing. No longer.

These days, Mr Marra is engaged in a cyber version of cat-and-mouse, frantically tracking the “bad guys” disseminating fake news – then trying to shut down their accounts.

“This work is adversarial – people are trying to penetrate our defences,” Mr Marra told an audience at the Aspen Ideas Festival last month.

Using language that might have emanated from the Wild West, he continued: “It sucks that we have to fight the bad guys… and the bad guys are creative. But we believe deeply in what we do…and in the fight.”

Should we feel reassured? After all, it’s become clear that fake news has circulated on a variety of social media platforms.

These range from the infamous “Pope Francis Shocks World, Endorses Donald Trump for President” headline, which originated on the now defunct fake news site WTOE 5 News and was shared almost a million times on Facebook, to mob violence and the killing of a woman last week in India after false rumours of child abduction were spread via WhatsApp.

Last week the US Senate Intelligence Committee supported three US intelligence agencies’ conclusion that Russia meddled in the 2016 US presidential election.

And revelations earlier this year that data-analytics group Cambridge Analytica harvested millions of Facebook profiles to target specific voters with fake news have left the social media company under investigation from no fewer than four US regulatory agencies – a situation that has not been helped by Facebook’s tardy response to tackling the problem.

Finally, though, Facebook is starting to fight back.

No, company executives are not (yet) admitting liability for what appears on their platform; nor are they volunteering to subject themselves to stricter regulation.

But they are trying to be a little more transparent. Hence Mr Marra (and his Facebook colleagues) turning up in Aspen to speak about their fight against “the bad guys”.

According to Mr Marra and Campbell Brown, global head of news partnerships at Facebook (and herself a former news anchor), one way that the company is engaging in this fight is by redeploying dozens of engineers to the task.

Instead of scrambling to create innovative products (such as a “stickier” News Feed), some of its best coders are now developing tools to track and eliminate fake news.

Just last week, for example, Facebook acquired London-based start-up Bloomsbury AI, which has developed an artificial intelligence that “reads text documents and answers questions about their contents”.

The company is also trying to develop fact-checking capability by hiring “news publisher specialists” and collaborating with groups such as the Associated Press, FactCheck.org and PolitiFact to fact-check fake news.

When it is identified, engineers either remove it completely or try to prevent it going viral by informing users that it is fake.

“Once we get [a story] fact-checked, we can reduce viewing of fake news by up to 80 per cent,” Mr Marra said.

This all sounds eminently sensible but there are many problems.

It is not always easy to determine what is fake or to remove political posts without seeming to engage in censorship.

It is also hard to fact-check the news fast enough, before stories go viral.

Nor, despite Mr Marra’s confidence, is it always easy to persuade users to stop looking at fake news: when Facebook started attaching “Disputed” flags under some articles last year, it found they actually reinforced beliefs, and so the company ditched them.

The collaboration with independent fact-checking groups has also been challenging.

“Both Facebook and the news organisations want to improve the quality of online media,” observed a recent report by Mike Ananny for the Tow Center for Digital Journalism.

“[But] there is an ongoing struggle within the partnership to define ‘fake news’ in a way that doesn’t leave most of the classification power with Facebook” – even as the social media giant keeps denying that it is a publisher.

In particular, the fact-checking groups told Annany that they felt frustrated that they did not know exactly how Facebook was using their fact-checks, and expressed “a general unease about how opaque and unaccountable much of the arrangement is”.

Another, more surprising issue is the source of fake news itself.

Most might assume that it comes largely from political groups or malevolent governments.

Not so, according to Mr Marra: he says that “90 per cent of fake news is driven by financial incentives”, ie, from commercial entities that use salacious stories to create advertising clicks on websites, such as those selling diet pills.

“The bad guys just want a click, they want an ad dollar – their incentives rely on you clicking from News Feed over to their site, so they will create sensationalist headlines,” Mr Marra says.

Maybe this is less alarming than the idea that America is being attacked by an army of Russian bots. Or maybe not – given the sheer size of the ad-tech industry and the dominance of consumer culture in the US.

Either way, one thing is clear: Facebook’s self-styled fight with the “bad guys” is likely to become more contentious as the midterm elections approach. THE FINANCIAL TIMES

ABOUT THE AUTHOR:

Gillian Tett is US managing editor at The Financial Times.