Facebook Inc. announced today that it will go ahead with a previously considered plan to reduce the amount of political content people see on the platform.

Talking on the quarterly earnings call late last month, Chief Executive Mark Zuckerberg said that although he doesn’t want to ban politics-related talk on Facebook, he wants to “turn down the temperature and discourage divisive conversations.” He didn’t, however, say how that might be done or where it will happen.

Aastha Gupta, a Facebook product management director, let that be known today. The first countries to experience a reduction of political content will be Brazil, Canada and Indonesia. That will happen this week, with the U.S. getting the same treatment in the coming weeks.

“During these initial tests we’ll explore a variety of ways to rank political content in people’s feeds using different signals, and then decide on the approaches we’ll use going forward,” said Gupta. He added that COVID-19-related content won’t be affected as long as it’s not coming from dubious sources – something Facebook has just cracked down on.

The initiative won’t mean people’s news needs will be completely shorn of political content. Instead, the algorithm will do some pruning to ensure such content doesn’t often appear at the tops of news feeds. “Our goal is to preserve the ability for people to find and interact with political content on Facebook, while respecting each person’s appetite for it at the top of their News Feed,” said Gupta.

Last month, Zuckerberg said Facebook had spent some time interacting with users to try to ascertain what experiences they wanted on Facebook. The conclusion, according to Zuckerberg, was a lot of people are fed up with the divisiveness that comes with political fighting.

As for how it works, Facebook has said it will use machine learning technology to predict if a post is political or not. It won’t matter if the content is posted by a friend or a media company or anyone else. A Facebook spokesperson told the New York Times that this is a work in progress, saying that the company is “refining this model during the test period to better identify political content, and we may or may not end up using this method longer-term.”

Photo: Zsolt Palatinus/Flickr