Facebook said it is making steady progress in curbing hate speech and violence on the social network, AFP reported.

The social network claims that improvements in artificial intelligence help it filter out hateful and aggressive content, as well as, among other things, bullying, inappropriate sexually explicit material, and fake accounts.

According to Facebook, the prevalence of hate speech in the first quarter of 2021 ranged from 0.05 to 0.06 percent, which is almost half of what it was in mid-2020.

Facebook VP Guy Rosen told reporters that the use of this prevalence metric was appropriate because it reflects not what we did, but what we missed. The prevalence of hate speech on Facebook continues to decline.

Violent and graphic content has dropped to less than 0.04 percent on Facebook and less than 0.02 percent on Instagram, and the company has also reduced the prevalence of other inappropriate content, including nudity and sexually explicit material.

The report said that Facebook removed 25.2 million pieces of hateful content, 8.8 million pieces of bullying and harassment content, and 9.8 million posts, including organized hate content, in the quarter.

Facebook said it is committed to enforcing a policy banning malicious misinformation about Covid-19 and has removed more than 18 million pieces of similar content from its platform and Instagram worldwide since the start of the pandemic.

According to the company, improvements in automated technology have allowed Facebook to make progress in identifying inappropriate content, not only in text but also in images, videos, or even memes that combine these elements.