« Back to Posts

Week In Review: Facebook Needlessly Changes Interface: Trending Stories Is The Real Source Of Fake News; AI Confirms Its Place As Year’s Most Overused Tech Term

1 Facebook Needlessly Changes Interface: Trending Stories Is The Real Source Of Fake News

Like a modern-day Dr. Frankenstein, Zuck is starting to realize just what a monster he’s created and is taking steps to ameliorate things.

The latest is an announcement that the News Feed is being purged of publisher stories and will rely instead on stories posted by friends, with “engagement”—the number of Likes and comments—being the key metric.

Why It Matters

This is all well and good only I don’t really find the number of publisher stories in my feed to be a problem. Almost all of what’s there has been posted by the half dozen or so friends I have who frequently post articles; the others, the ones who post clickbait stories ranging from felines to politics, have all been put on “ignore.”  If anything, I find there’s less content from publishers I want to hear from, publishers whose Facebook pages I have actually gone and Liked.

Where I do find #FakeNews however, is when I click on the “Trending” stories section in the right hand column of the desktop app. Clicking on one of those stories will turn up on odd mix of mainstream news sources (ESPN, Buzzfeed, The BBC) and a slew of random sites I’ve never heard of.

Many of those sites linked to from the Trending section are all about the #FakeNews, something Buzzfeed’s Craig Silverman has been very diligent about pointing out.  In fact, I became aware of it when a “Trending” link I clicked on in October 2017 turned up a clearly inflammatory headline about nefarious “Zionist billionaires” and a google search turned up a November 2016 Buzzfeed story calling out the site responsible for that headline (which, FWIW, is run out of Minnesota, not Macedonia) as being a major source of #FakeNews on Facebook.

And yet a year later, after both Zuck and Sheryl had promised to make ending #FakeNews a priority, the story still wound up at the top of the Trending News section.

Physician heal thyself and all that. The problem isn’t the News Feed, it’s the Trending Stories section, as it doesn’t take much to figure out that the odds of users sharing those inflammatory #FakeNews stories to their News Feed is extremely high.

Or to put it in tweetable form: Wake up FB! #FakeNews in your Trending section is what infects the News Feed. That’s what you need to fix.

Why It Matters For Publishers

A very apt tweet from The Verge’s Silicon Valley editor, Casey Newton, sums up the issue for publishers: “So many publishers think they have audiences, when what they really have is traffic. I think we’re about to find out who has an audience.”

This is actually great news for TV networks though, because compared to most online publishers, they have a massive audience. What’s more, they have fans, people who want to hear from them, or at the very least, from one of their shows.

What You Need To Do About It

If you’re a web publisher who relies on Facebook traffic, you need to rethink your entire business model. That sucks, but clickbait headlines that seem to embody the concept of A/B testing are no longer going to work. You need people who like you for who you are. Not who you’re pretending to be. (And other advice from ABC Afterschool Specials.) A paid campaign to get your new and improved self in front of people is probably in order too. Just wait until you’ve got work you’re proud of.

If you’re a television network, high five! You have an audience and people who like you, now’s your chance to double down on that. You may want to use this opportunity to run some ads on Facebook with the express goal of adding Likes to your page (or the pages of your individual shows.) But right now it seems like your posts are going to appear in the News Feeds of people who want to hear from you, something many digital publishers can no longer count on (unless, of course, they pay for it.)

If you’re an advertiser… guess what? You’re fucked again! Sort of. What seems to be happening is that you’ll need to restrategize. The new Facebook will likely have fewer ads you’ll need to pay more money for. Which is not a bad thing in the long run, as ads are more effective when they are few and far between.


2. “AI” Confirms It’s Place As Year’s Most Overused Tech Term

Among the many interesting things to come out of CES this year, is confirmation that “AI” or “Artificial Intelligence” continues its reign as the world’s most overused tech term.

As we mentioned earlier this week, in the hands of a good copywriter, any sort of predictive technology becomes “AI.”  This includes any sort of “If This Then That” (IFTTT) functionality, so that a system that is automatically programmed to ask if you want to record all episodes of a show or just one, magically becomes AI.

Why It Matters

To be fair, this sort of functionality is a low-level version of AI, and most of the time actually does improve the functionality and usefulness of the product. What differentiates it from true AI is that it’s based on prior behavior—the UX team has discovered that asking people if they want a text reminder that their favorite show is about to start generally results in higher tune-in, so that becomes an automatic next step once someone favorites a channel (If they favorite a channel, then ask them ….)

True AI is far more advanced. It uses human level reasoning to sort through thousands of variables and draws conclusions using advanced calculations. Think of it as a much more intelligent, personalized and intuitive version of IFTTT-style AI.

Recommendations are one area where advanced AI can really shine, analyzing a viewer’s prior habits, likes, dislikes, demographics and purchases to intuit that, say, they liked Pretty Little Liars because they like series set in beachfront towns with less than ten episodes, not because they like Reece Witherspoon or psychological dramas. (And as TV moves from time-based interfaces to library-based ones, recommendations will become more and more important.)

What You Need To Do About It

Not much to do, beyond understanding the difference between basic AI and more advanced forms, with an eye towards not lumping them all together and recognizing the multiple advantages of the latter.