Hate speech on the decline on Facebook, company reports
Hate speech is on the decline on Facebook, according to the company’s latest Community Standards Enforcement Report released last week.
The social media company removed 31.5 million pieces of hate speech content the site in Q2, compared with 25.2 million in Q1, as the the prevalence of hate speech on the site sees its third quarterly decline.
According to the report, Facebook removed 6.2 million pieces of organised hate content in Q2, compared to 9.8 million pieces in Q1, as well as 34.1 million pieces of violent and graphic content, up from 30.1 million in Q1.
In Q2, hate speech was appearing in five posts per 10,000, while it was present in five or six per 10,000 in Q1.
These improvements come following investment in AI that better enable Facebook to detect various hate speech violations. It said steady, continuous AI improvements and advancements, such as the Reinforcement Integrity Optimiser (RIO), help its AI models to spot hate speech using real-world data and improve over time. It also said changes to the News Feed had a positive impact.
With these changes, more than 90% of the content Facebook took down across 12 of its 13 policy areas was removed before being reported by a user.
Since Facebook began reporting on the issue, its removal of hate speech content has increased 15-fold across both Facebook and Instagram.
Meanwhile, Facebook emphasised its commitment to reducing misinformation around Covid-19 on the platform. Since the start of the pandemic, it has removed more than 20 million pieces of content from Facebook and Instagram globally for violating its policies. It removed 3,000 accounts, pages, and groups for repeatedly violating our rules against spreading Covid-19 and vaccine misinformation.
In terms of child safety, Facebook removed 2.3 million pieces of child nudity and physical abuse content from Facebook and 458,000 from Instagram, as well as 25.7 million instances of child sexual exploitation content from Facebook and 1.4 million from Instagram.
There was also a significant increase in the number of suicide and self-injury posts being removed from the site – 16.8 million, up from 5.1 million in Q1. Facebook said this was due to a technical fix which allowed it to go back and catch violating content it had previously missed.