To curb hate speech, Facebook removes 2 crore posts in three months

In an attempt to curb hate speech, Facebook removed a whopping 2.25 crore posts from its platform during the April-June quarter, using both artificial intelligence and manual reviews, after receiving flak from several advertisers for alleged inaction in this regard.

The social media giant stepped up proactive and suo moto actions during this period, removing content that violated their community standards, even before they could be reported by users.

According to the figures published in the sixth edition of Facebook’s Community Standards Enforcement Report, its first quarterly update, as many as 2.25 crore posts were found to be “hate speech” and removed. In the January-March period, Facebook had taken 96 lakh such actions.

Facebook defines ‘hate speech’ as “violent or dehumanising speech, statements of inferiority, calls for exclusion or segregation based on protected characteristics, or slurs.” These characteristics include race, ethnicity, national origin, religious affiliation, sexual orientation, caste, sex, gender, gender identity, and serious disability or disease.

The report further gave an account of the actions taken on its photo and video sharing platform, Instagram. Around 33 lakh posts were removed from the platform during the April-June period for violating its community policy, as compared to 8 lakh in the previous quarter.

Facebook used artificial intelligence (AI) to detect these posts more than human reviewers, the report said. It said, though technological intervention has helped much during the pandemic period, there are certain areas where it relies on people for reviewing its content. The number of human reviewers had declined due to the pandemic.

The company pursued a proactive policy in its attempt at curbing hate speech on its platform. Of the total actions taken this quarter, 94.5 per cent were suo moto, as compared 88.8 per cent in the previous three months. It followed a similar policy for Instagram; 84.2 per cent of its actions were suo moto as compared to 44.5 in the previous quarter.

The company also mentioned that it restored around 1.52 lakh posts on Facebook and around 9,300 posts on Instagram after scrutiny, as they were found to have been removed incorrectly.

The expansion of automation technology in English, Spanish, Arabian, Burmese, and other languages have helped the platform perform better on the content violation front.

Facebook had recently faced flak from major advertisers due to alleged inaction against posts containing hate speech. Many giant advertisers like Ford, Verizon, Coca-Cola had joined the movement, #StopHateForProfit in July. Microsoft, Pepsi, and Starbucks had even taken their ads off Facebook during this time.

Related news: When social media runs riot, and paves the mob way