FB ramps up Sinhala and Tamil language experts to check hate speech
Facebook has significantly ramped up Sinhala and Tamil language expertise to curb hate speech and other harmful contents originating from Sri Lanka on the social media platform after mob violence directed at its Muslim minority post the Easter Sunday bombs, which saw hate speech and rumours spreading like wildfire on social media services, especially those operated by Facebook.
This led to the company hosting a community standards briefing session for media organisations in Colombo on Wednesday to increase understanding and awareness of Facebook’s Community Standards, and the work they do to keep the community safe.
This also came after the government briefly shut citizen’s access to social media networks. The Easter Sunday attackers had preached extremism and hate speech on Facebook leading up to the bombings, which killed over 250 persons, and wounded over 500. As Facebook grapples with the spread of hate speech (which includes content that directly attacks people based on their race, religion, national origin, gender identity, caste, sex, ethnicity, serious disability and sexual orientation) on its platform, it is introducing changes that limit the spread of messages.
“We have invested heavily on Sinhala and Tamil language experts as we have promised. We significantly ramped up our language expertise in Sri Lanka through hiring more experts after the Digana (Kandy) riots,” a Facebook spokesperson told reporters in Colombo at the event. The Sinhalese extremist leader Amith Weerasinghe who was arrested following last year’s riots had managed a Facebook page with over 150,000 followers, which was taken down only after the violence began. Similarly, music artist Iraj Weeraratne’s Facebook page was suspended on July 7 this year.
At a media briefing on Wednesday he claimed that the decision to remove his Facebook page was motivated by political factors. But the Business Times reliably learnt that he had repeatedly violated FB rules, which is why his page was removed.
In a statement a Facebook spokesperson told the Business Times, “As per our Community Standards, we do not allow any organisation or individual that engages in hate, violence or terrorist activity. Content that expresses support or praise for individuals or groups engaged in organised hate, violence or terrorist activity has no place on our platform and, as per our Community Standards, we will remove content that expresses support or praise for individuals or groups involved in these activities. Repeated violations of our policy may lead to a Page being unpublished.”
According to Facebook’s latest community standards enforcement report, the social media platform removed four million pieces of hate speech globally from the platform in the first quarter of the year and took action on 33.6 million pieces of violent content of which 98 per cent was once again identified by FB technology. The official said 2.2 billion fake accounts have been closed while 1.8 billion pieces of spam were removed.
The spokesperson on Wednesday said that 65.4 per cent of hate speech content was removed by its proactive detection tools, compared to 51.5 per cent in 3Q18, which is a significant improvement. “We have invested in proactive detection tools; AI tools become stronger over the last few years,” he said noting that Facebook as hate speech is highly contextual, FB has to rely on the community to report such content.
Facebook is in the process of developing a global metric while expanding prevalence measurement to cover more languages and regions, to account for cultural context and nuances for individual languages.
Facebook has deployed a team of 30,000 to work on safety and security issues of its users globally, of which over 15,000 have been deployed to look into content-specific issues.