Facebook a ‘megaphone for hate’ against India’s minorities

Campaign group accuses Facebook of failing to rein in hate speech inflaming ethnic tensions in India’s Assam.

facebook logo
Avaaz said Facebook was too reliant on artificial intelligence hate speech detection mechanisms [File: Reuters]

Facebook is failing to rein in a “tsunami” of hate posts inflaming ethnic tensions in India’s Assam state, campaign group Avaaz said on Wednesday, with content about “criminals”, “rapists”, “terrorists” and “dogs” shared almost 100,000 times.

Avaaz said the dehumanising language – often targeting India‘s Bengali Muslims – was similar to that used on Facebook about Myanmar’s Rohingya before an army crackdown and ethnic violence forced 700,000 Rohingya to flee in 2017 to Bangladesh.

A Facebook spokesman said the company was proactively removing 65 percent of hate speech before it was reported globally, and that its content reviewers, working in at least nine Indian languages, were rigorously trained to “catch harmful content”.

The United Nations warned of rising hate speech on social media in India’s northeastern Assam state as the government prepared to publish a list of Indian citizens in the state that left out nearly two million people, many Bengali Muslims.

191003101543616

“Facebook is being used as a megaphone for hate, pointed directly at vulnerable minorities in Assam, many of whom could be made stateless within months,” Alaphia Zoyab, a campaigner at Avaaz, said in a statement.

To be included on the register, published in August, residents had to prove their families lived in India before March 24, 1971, when hundreds of thousands of people began fleeing conflict across the border in what is now Bangladesh.

Blaming outsiders

Resentment against undocumented immigrants has simmered for years in Assam, one of India’s poorest states, with residents blaming outsiders, mainly Bangladeshis, for taking their jobs and land.

Avaaz, which campaigns against disinformation on social media, analysed 800 Facebook posts and comments relating to Assam and the register, and found that about 27 percent constituted hate speech, which it estimated had been viewed five million times.

It flagged 213 posts to Facebook, which removed 96 of them, including one by an elected official implying that those who rape “our mothers and sisters” are “Bangladeshi Muslims”, and calls to “poison” daughters and legalise female foeticide.

190914153846262

Facebook has been criticised for not acting to stamp out hate speech on its platform in Myanmar and for not stopping the use of fake accounts aiming to sway public opinion in the 2016 US election and the UK’s vote to leave the European Union.

Avaaz said Facebook was too reliant on artificial intelligence hate speech detection mechanisms, and needed to put in place more human-led teams to monitor content in Assam.

“Despite the clear and present danger faced by these people, Facebook is refusing to dedicate the resources required to keep them safe,” said Zoyab.

“Through its inaction, Facebook is complicit in the persecution of some of the world’s most vulnerable people.”

Facebook’s spokesman said the use of artificial intelligence had improved its hate speech detection and that it had invested in tools to encourage people to flag up inappropriate content.

“While we’re making progress, our technology isn’t perfect yet. That’s why reports from our community are still incredibly important,” he said.

Source: Reuters