Facebook is hesitant to curb disagreeing user content in India

New Delhi, India (Associated Press)- Facebook According to leaked documents obtained by the Associated Press, India has been selectively curbing hate speech, misinformation and inflammatory posts, especially anti-Muslim content, even though its employees have expressed doubts about the company’s motives and interests.

From the study in March this year to the company memorandum dating back to 2019, internal company documents concerning India emphasized FacebookIn the world’s largest democracies and companies’ largest growth markets, it has been working hard to eliminate abusive content on its platform. Community and religious tensions in India have always boiled over social media and incited violence.

File display Facebook I have been aware of these problems for many years and have raised questions about whether they have taken sufficient measures to solve these problems. Many critics and digital experts say it has failed to do so, especially with the participation of members of the Bharatiya Janata Party (BJP) governed by Prime Minister Narendra Modi.

all around the world, Facebook It is becoming more and more important in politics, and India is no exception.

Modi was praised for using the platform to gain party advantage during the election, and the Wall Street Journal’s report last year was suspicious. Facebook Selectively implement its hate speech policy to avoid counterattacks from the BJP.Modi and Facebook Chairman and CEO Mark Zuckerberg exudes a cordial atmosphere. In 2015, the two Facebook Headquarters.

The leaked documents included numerous internal company reports on hate speech and misinformation in India. In some cases, its own “recommendation” features and algorithms have enhanced most of them. But they also include company employees’ concerns about improper handling of these issues and their dissatisfaction with viral “dissatisfaction” expressed on the platform.

According to the document, Facebook View India as the “highest risk country” in the world, and regard Hindi and Bengali as a priority for “violating the automation of hostile speech.” However, Facebook There are not enough local language moderators or content tags to prevent misinformation that sometimes leads to real-world violence.

In a statement to the Associated Press, Facebook Said that it has “invested heavily in technology to discover hate speech in various languages ​​including Hindi and Bengali”, thereby “cutting the number of hate speech people see by half” in 2021.

“Hate speech against marginalized groups, including Muslims, is on the rise globally. Therefore, we are strengthening law enforcement and are committed to updating our policies as hate speech evolves online,” a company spokesperson said .

This Associated Press story and other stories being published are based on information disclosed to the Securities and Exchange Commission and provided to Congress in edited form by the predecessor. Facebook Legal counsel to the whistleblower Frances Haugen, who is an employee. The edited version was obtained by a consortium of news organizations including the Associated Press.

As early as February 2019 and before the general election, people’s worries about misinformation were high. Facebook If all they do is follow the pages and groups recommended by the platform itself, the employee wants to know what new users in that country see on their news feeds.

The employee created a test user account and kept it for three weeks, during which an extraordinary incident that shocked India occurred-a radical attack in controversial Kashmir killed more than 40 Indian soldiers. The country is close to a war with rival Pakistan.

In the note titled “An Indian test user caught in a sea of ​​polarized, nationalist information”, the employees whose names were edited said they were “shocked” by the content of the news source. Nationalist content, misinformation, violence And bloody. “

Recommended seemingly benign and harmless groups Facebook It quickly became something else entirely, and hate speech, unsubstantiated rumors, and viral content were rampant.

The recommended groups are full of fake news, anti-Pakistan rhetoric and Islamophobic content. Most of the content is very vivid.

One of the men was holding the bloody head of the other man, covered with the flag of Pakistan and the flag of India on it.It’s “popular in FacebookThe function displays a series of unconfirmed content related to India’s retaliatory attack on Pakistan after the bombing, including a napalm image in a video game clip, in which one person debunked FacebookOf fact-checking partners.

The researcher wrote: “After this test user’s news feed, I have seen more pictures of the dead in the past three weeks than I have seen in my life.”

It aroused people’s deep concern about the possible consequences of such fragmented content in the real world. At that time, local media reported the Kashmiris being attacked in the aftermath.

“As a company, should we take additional responsibility to prevent integrity damage caused by recommended content?” the researchers asked in their conclusions.

The memo distributed with other employees did not answer this question. But it does expose how the platform’s own algorithms or default settings can play a role in stimulating such dissatisfaction. The employee pointed out that there are obvious “blind spots”, especially “local language content”. They said they hope that these findings will trigger a conversation about how to avoid this “integrity damage”, especially for those who are “significantly different” from the typical American user.

Although this study was conducted in three weeks, which is not a general representation, they admit that it does show how this “unreviewed” and questionable content “completely takes over” during a “major crisis event” “.

this Facebook The spokesperson said that this test study “inspires deeper and more rigorous analysis of its recommendation system” and “helps product changes to improve them.”

“In addition, our work to curb hate speech is continuing, and we have further strengthened the hate classification to include four Indian languages,” the spokesperson said.

Sign up for the daily newsletter

Copyright © 2021 The Washington Times, LLC.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *