Facebook is improving its efficiency at recognizing and taking down hateful content. The platform's Community Standards Enforcement Report shows a significant reduction in the prevalence of hate speech in the last quarter of 2020.

Hate Speech on Facebook Has Dipped

In Facebook's Community Standards Enforcement Report for October through December 2020, the platform revealed major improvements in the way Facebook is handling hate speech. Guy Rosen, Facebook's vice president of integrity made a post on the About Facebook blog to highlight its success.

For every 10,000 views of content of Facebook, only seven to eight were views of hateful content. Facebook also took action on 6.3 million bullying and harassment posts, an increase from 2.5 million posts in quarter three.

The platform was also able to catch 6.4 million pieces of organized hate content, which is a massive improvement compared to the four million pieces of content that Facebook removed in the previous quarter.

The Community Standards Enforcement Report showed similar improvements on Instagram. Five million posts pertaining to harassment and bullying were removed on Instagram, along with 6.6 million pieces of content containing hate speech and 308,000 pieces of organized hate content. All of these numbers surpass the statistics recorded in quarter three.

Facebook's proactive rate, otherwise known as the amount of content Facebook removes before a user reports it, also improved. For bullying and harassment, the platform's proactive rate spiked from 26 percent to 49 percent on Facebook, and from 55 percent to 80 percent on Instagram.

Facebook's hate-detecting AI has played a major role in boosting its response to hate speech. Rosen notes that the platform has made improvements to its AI "where nuance and context are essential, such as hate speech or bullying and harassment," which ultimately helped reduce the amount of hateful content on Facebook.

Mike Schroepfer, Facebook's chief technology officer, made a separate post on the About Facebook blog to celebrate the achievements of Facebook's ever-improving AI. Despite its recent accomplishments, Schroepfer hopes that Facebook's AI can get better "at viewing content in context across languages, cultures, and geographies."

Facebook's AI Is Getting Smarter

Facebook began relying more on its AI system to help evaluate hateful content due to the ongoing COVID-19 pandemic. Rosen noted that the Facebook team will likely still be affected by the pandemic "until a vaccine is widely available."

Even though Facebook's moderators will one day return to work as normal, it doesn't seem like the platform's AI is going anywhere. Facebook has only made its AI smarter and more efficient, and it doesn't seem like Facebook plans to stop progressing anytime soon.