Get my Breaking World News by E-MAIL

SEARCH THIS BLOG.

Subscribe to my World News Updates

Translate

MY MUSIC CELEBRITY PHOTO ALBUM

Wednesday, August 31, 2016

Facebook FIRES Trending Team as algorithms were flawed on Megyn Kelly Trend #HNNTech

Image result for facebook fired

Now if I were in Lagos, Nigeria today, that would have been my question for Mark Zuckerberg. If you are a trending news website like #HNNAfrica you would have noticed there are no more details on trending stories on Facebook. Facebook is not a media company but it is the source of news for 44% of adults out there. What happened severely backfired and as much as 18 people were FIRED!

It’s been a rough few days for Facebook’s Trending Topics. The company announced on Friday that it’s made the news list more automated by removing article descriptions and leaning more heavily on algorithms to produce it. 

The company also reportedly fired as many as 18 editorial contractors responsible for writing descriptions and ensuring the accuracy of the sources it uses. Now we’re watching the immediate — and disastrous — effects unfold. Yesterday, a more serious error occurred: a fake news story about Fox News commentator Megyn Kelly’s supposed secret affinity for Hillary Clinton blew up on Facebook, landing it on top of the Trending list. Not only did the engineers (or algorithms) responsible for Trending fail to realize the story was false — it came from a partisan libertarian source called End the Fed — but Facebook also left it in the Trending module for hours to collect likes and comments. All the while, the company unwittingly gave enormous exposure to a damaging piece of false information.

So over the course of a single weekend, Facebook’s attempts to reorient one of its most prominent news products backfired in a big way. In its efforts to rely more heavily on algorithms, the company exposed a central vulnerability in its approach to news. All it took was 72 hours, and a false news story in the heat of a polarizing US election season, to show how fallible its algorithms can be.

Blog Archive