Facebook takes on 'fake news' with AI in the new age of 'alternative facts'

Profile picture for user slauchlan By Stuart Lauchlan February 1, 2017
In the new world order where 'alternative facts' is apparently a thing, Facebook CEO Mark Zuckerberg is tackling the 'fake news' controversy with AI and algorithms.

Mark Zuckerberg
Facebook had a bad day in court yesterday, landed with a $500 million fine. But at the same time, it turned in fourth quarter sales of $8.8 billion, up 51% year-on-year, as user numbers head towards the two billion mark. So it’s no wonder COO Sheryl Sandberg was able to describe the fine as “non-material”.

The court case centered on allegations by games publisher ZeniMax Media that employees  of Oculus, the virtual reality firm acquired by Facebook for $2 bililon, stole ZeniMax trade secrets. The jury found in favor of ZeniMax Media, but didn’t support its claim for $4 billion in recompense.

The ruling didn’t dent Wall Street’s enthusiasm for Facebook, whose share price rose on the back of stellar full year numbers. Revenue was up 54% year-on-year to $27.6 billion, while net income soared from $3.7 billion dollars to $10.2 billion dollars.

There was the usual dazzling array of metrics to be absorbed:

  • 1.15 billion people accessed Facebook on mobile on an average day in December, up 23% year-on-year.
  • Mobile ad revenue reached $7.2 billion, up 61% year-on-year, and is now approximately 84% of total ad revenue.
  • Instagram now has over 600 million monthly actives and recently passed 400 million daily actives.
  • Instagram Stories reached 150 million daily actives just five months after the launch.
  • 400 million people now use voice and video chat on Messenger every month.
  • More than 50 billion messages are sent through WhatsApp every day.
  • WhatsApp has passed 1.2 billion monthly actives.

Sandberg cited the growing use of video content for advertising and marketing purposes as a driver of growth, but also as a sector of the market in which Facebook can help to shape its evolution:

One of the things that's really important in this is helping marketers understand that they need to optimize those video ads. An ad that works really well 30 seconds in other platforms and more traditional platforms can work on ours, but the ones that are optimized and use our targeting really perform better. We're working hard with advertisers to help them see that.

So to share one example, Motorola, working with Ogilvy and Moto Mento, launched the Moto Z phone. They did awareness boosting before they launched, targeting Android users and Verizon subscribers. They optimized their video for the Facebook and Instagram mobile feeds. Then after they launched, they did purchasing ads and re-targeted people who had viewed those initial ads.

That's just a great example of someone using video ads, optimizing a format, but also using the pretty unique targeting we can offer to drive sales. They measured that they had over a 3.5% lift in sales driven by the Facebook and Instagram video ads…We're going to work client by client to get the video format of those ads right, get the targeting to be as good and as deep as it can be and make sure we're measuring all the way through to sales.


Away from the stats, CEO Mark Zuckerberg faced down the ‘fake news’ controversy that arose during the US Presidential Election campaign and has rumbled on into the new age of ‘alternative facts’:

We see Facebook as a community, and ourselves, our role, as supporting that community. We don't write the news that you read, but we want to be a place where people can access information and have meaningful conversations, and this is a responsibility that we take very seriously.

In the past, we've taken steps to reduce spam and clickbait, and now we're approaching misinformation and hoaxes the same way. In Q4, we started working with third-party fact-checkers in the U.S. to flag disputed stories and make them less likely to appear in News Feed. We've made it easier to report and identify misinformation and we're working to build stronger ties between Facebook and the news industry.

Our primary goal here is to do the right thing for our community. If we can help people stay informed and make Facebook a better place to understand what's going on in the world, then we think that's going to make our community stronger and a more positive force for good in the world.

The firm has added a ranking algorithm that identifies misleading, sensational or spammy posts and lowers them in the News Feed as part of an update. The update, according to a Facebook news release, is to support “authentic communication” in the News Feed: 

When ranking News Feed we look at many signals personal to you, such as how close you are to the person or Page posting, as well as more universal signals like the overall engagement (likes, comments, shares) that a post has.

With this update, we’re adding new universal signals to determine whether a post might be authentic. To do this, we categorized Pages to identify whether or not they were posting spam or trying to game feed by doing things like asking for likes, comments or shares. We then used posts from these Pages to train a model that continuously identifies whether posts from other Pages are likely to be authentic. For example, if Page posts are often being hidden by people reading them, that’s a signal that it might not be authentic.

If a post is likely to be authentic based on the new signals we look at, it might show up higher in your feed.

Another way of tackling the ‘fake news’ phenomenon might be to use AI tech, Zuckerberg suggested:

There's an increasing focus on objectionable content, right, and a lot of unfortunate things, right, that people share on Facebook. It's a minority of the content, but I'm really focused on making sure that our company gets faster at taking the bad stuff down. We can do better with people, but ultimately the best thing that we can do is build AI systems that can watch a video and understand that it's going to be problematic and violate the policies of our community and that people aren't going to want to see it and then just not show it to people before bad experiences happen and things like violent content gets spread through the network.

AI is both going to be great on showing people content that's really good and helping us enforce the community standards that we have to make sure that everyone has a good and fair experience.

My take

Just remarkable numbers, of course. The ‘fake news’ initiatives are welcome, taking the fight proactively to critics, although in the new world order where ‘alternative facts’ are presented as a perfectly reasonable reality, Facebook can expect to continue to come under fire from all sides.