After training its News Feed algorithm to recognize clickbait headlines in 2016 and penalized the websites and Pages associated with those posts, Facebook is getting more precise in its fight.
In May 17th, 2017, the social giant said that it can now target and pinpoint posts that link to those articles.
Previously, Facebook was able to track website domains or Facebook Pages at large when hunting for clickbait. While this enables Facebook to be more precise as it can target individuals, but it also made it harder for it to quarantine the occasional clickbait from an otherwise reputable publisher.
By improving its algorithms, Facebook can target a lot more precisely as it can strike down to posts.
Facebook expects that most Pages "won’t see any significant changes" to their reach. Publishers that are affected, won't be be further impacted if they stop posting clickbaiting posts, according to the company.
"People tell us they don't like stories that are misleading, sensational or spammy," wrote Facebook engineers Arun Babu, Annie Liu, and Jordan Zhang in a blog post. "That includes clickbait headlines that are designed to get attention and lure visitors into clicking on a link. In an effort to support an informed community, we're always working to determine what stories might have clickbait headlines so we can show them less often."
To target clickbaits that are common for spreading fake news and hoaxes, Facebook said that it is not only analyzing the bulk posts of a page. The algorithms will also look at two distinct signals: whether a headline "withholds information or if it exaggerates information separately."
This strategy is supposed to make Facebook more effective when evaluating posts.
To do this, the company uses Artificial Intelligence to train its algorithms. By feeding the computers a bunch of headlines, the algorithms can understand what a clickbait headline looks like.
"We categorized hundreds of thousands of headlines as clickbait or not clickbait by considering if the headline exaggerates the details of a story, and separately if the headline withholds information. A team at Facebook reviewed thousands of headlines using these criteria, validating each other’s work to identify large sets of clickbait headlines," said Facebook on its blog post.
its effort is has also improved because the algorithms look at clickbait-detecting signals separately instead of trying to determine if a post or page is guilty of a number of joint factors. It does this similarly to how email spam filters. It then automates this process on the News Feed by looking for commonly used phrases that match the clickbait criteria.
Initially testing the algorithms in English, Facebook is starting to release it to other languages, evaluating posts that are written in Arabic, Chinese, French, German, Italian, Portuguese, Spanish, Thai and Vietnamese.