Over the Lok Sabha election period last summer, one in every seven tweets mentioning women politicians in India were “problematic” or “abusive”, amounting to over 100 such tweets to each woman politician every day, according to an Amnesty International India study of thousands of tweets.
Non-BJP women politicians faced 56.7 per cent more online abuse than women politicians from the BJP, according to the study.
As part of a global quantitative analysis on women contesting elections, the study also found that Indian women politicians faced far more online abuse (13.8 per cent of tweets) than their counterparts in the UK or the US on Twitter (7.1 per cent). Muslim women received almost 55.5 per cent more problematic or abusive content than women from other religions, it found.
Examples of sexist abuses from India included Hindi slurs for ‘witch’ and ‘prostitute’ as well as threats to send the politician concerned to Pakistan, the study found.
Given the significance of Twitter among the political elite, the report suggests, “Twitter is failing in its responsibility to respect women’s rights online.” It recommends more focus on regional languages in India, continuously evaluating its efforts against online violence against women, and more transparency about its content moderation process.
Amnesty showed their findings to Twitter in November, the report states. In a lengthy response, the company said their proactive technology solution increased suspended or locked accounts by 105 per cent.
The data set involved 1,14,716 random samples of 7 million total tweets mentioning 95 Indian women politicians between March and May 2019. Based on nomination papers, the sample of women were MPs in the two most recent Lok Sabha and Rajya Sabha elections, MLAs as of February 2019, party office-bearers and spokeswomen, current and former chief ministers, and members from reserved constituencies.
The categories of online abuse include threats of physical or sexual violence, caste or religious slurs, and sexist discrimination, the report mentions. The Decoders read tweets in English, Hindi, Bengali, Gujarati, Kannada, Malayalam, Marathi, Tamil and Telugu.
Between July and November 2019, the organisation crowdsourced 1,907 digital volunteers, who were given a random set of tweets. The decoder marked tweets as “abusive” or “problematic” as well as the nature of abuse. Each tweet was marked by multiple people, amounting to 4,74,383 total answers.
The data was then submitted to a team of data scientists “to validate and analyse”. Amnesty said it conducted this analysis in the US and the UK in 2018 to build the world’s “largest crowd-sourced dataset of online abuse against women”.
“Abuse, harassment and hateful conduct have no place on Twitter and we have taken strong steps to proactively address the health of the conversation on our service – including around peak moments such as Loksabha 2019. Today more than 50% of abusive content that we take action on is identified proactively using technology, instead of relying on reports from people using Twitter. Our work will never be done, and our product, policy and engineering teams continue to work at scale and pace to build a healthier Twitter,” a Twitter spokesperson said.
📣 The Indian Express is now on Telegram. Click here to join our channel (@indianexpress) and stay updated with the latest headlines