News

Facebook Whistle-Blower’s Complaint Details Importance Of Indian Polls, Duplicate Accounts, Hate Speech

An allegation has been filed© with the US Securities and Exchange Commission (SEC) by Whistleblower Frances Haugen alleging that Online accounts supposedly run by or affiliated with both the Rashtriya Swayamsevak Sangh (RSS), the ruling Bharatiya Janata Party's ideological fountainhead, promoted allegedly fear-mongering and dehumanising content (BJP).

According to the legal lawsuit with the SEC, “RSS (Indian nationalist organisation Rashtriya Swayamswvak Sangh) Users, Groups, and Pages promote fear-mongering, anti-Muslim narratives targeting pro-Hindu communities with V&I (violence and inciting) intent.”

An "Adversarial Harmful Networks – India Case Study" report submitted with the dispute cites "political considerations," the classification of India in "Tier 0" alongside the US and Brazil whenever it comes to the corporation's "Top 3 Political Preferences," and the inability of the corporation to address the problem of a subject matter against a specific religion in India.


Also Read: Global Outage & Whistleblower Revelations: Facebook Fights Duel Crises


According to our analysis, automatic checks remove only 0.2 per cent of reported hate speech. There is also the well-known issue of a shortage of language classifications that can also test for translations. According to the claim, Facebook's internal records demonstrate that the absence of Hindi and Bengali categories means that most of the submitted content, especially anti-Muslim narratives, was not handled with or identified by the algorithms.

India has around 340 million Facebook members, making it a significant portion of Facebook's 2.89 billion regular monthly visitors worldwide.

Hate speech in Facebook posts is detected using classifiers, which are automated systems and techniques.

Facebook has claimed that it is now identifying and removing more hate speech than it has ever been. It started in 2019 that the site's hate speech detection algorithms supported four languages of  India: Hindi, Bengali, Urdu, and Tamil.

Despite this, Haugen claims that Bengali classifiers are inaccessible in India for content monitoring.

Haugen's protest underscores the categorization that India receives in conjunction with the presidential elections in the United States and Brazil. Tier O comprises Brazil, India, and the United States, while Tier 1 includes Germany, Indonesia, Iran, Israel, and Italy, according to Facebook's "Top 3 Policy Priorities" for 2020. Nobody knows what the Tier 0 priority signifies in regards to Facebook's expansion in India.

In India, Facebook has been accused of taking no action in response to the content shared by some groups. The year before, the Wall Street Journal (WSJ) referenced an interim document in reporting that the Hindu nationalist group Bajrang Dal was a "threatening" organisation. However, due to financial and safety concerns, the corporation did not act on the information. According to the Wall Street Journal, Facebook "refused to remove the group" because its security team advised that taking action against Bajrang Dal "may jeopardise both the company's business prospects and its Indian employees. No judgment was made in the end regarding this matter.

Facebook's management of similar accounts, dubbed SUMA internally, is also being scrutinised. According to the information provided by Haugen, Facebook is doing very little to address the issue. Facebook was cognizant of the use of SUMAs by affiliates of India's ruling political party to magnify their narrative, according to an internal memo included with the case titled Lotus Mahal. According to Haugen's lawsuit, "a BJP IT Cell staffer coordinated message instructions to supporters with a copy-paste campaign targeting electorally themes."