Advertisement

Twitter accused of failing to protect women from online abuse by Amnesty International

Moran / Unsplash
Moran / Unsplash

One of the world’s most popular social media platforms isn’t doing enough to protect women from online abuse and violence despite promises to do so, according to new research by Amnesty International.

Back in 2018, the charity published a study in partnership with Element AI which showed that women are abused on Twitter every 30 seconds. The Troll Patrol study analysed tweets sent to female politicians and journalists in the UK and US. Another report, Toxic Twitter, demonstrated the overwhelming scale of the problem on the platform.

Two years on, Amnesty is again challenging the company for making little change to rectify the issues. Following the initial Toxic Twitter report, the charity made 10 recommendations on how to improve the situation. Since then, it appears Twitter has implemented only one recommendation, which is increasing transparency on how it handles reports of abuse.

Speaking about the new research, Rasha Abdul Rahim, co-director of Amnesty Tech, said: “Our analysis shows that despite some progress, Twitter is not doing enough to protect women users, leading many women to silence or censor themselves on the platform. We have outlined clear, straightforward steps that Twitter can take to make its platform a safer place for women to express their views. Twitter can and must do more to protect women from abuse.”

Part of the criticisms Amnesty is levelling at the social media platform is to do with the fact Twitter isn’t very transparent about its inner workings. In particular, Amnesty says it’s difficult to gauge the full extent of the problem of online abuse as the company doesn’t provide any meaningful data, such as country-level breakdowns on the number of user reports of abuse, or the specific kinds of abusive language.

In response, Twitter said if it published disaggregated data by country or region, this could be open to misinterpretation or give a “misleading impression” of the problem.

Previous Amnesty research found that former shadow home secretary Diane Abbott was the most abused female MP online during the 2017 General Election (Getty )
Previous Amnesty research found that former shadow home secretary Diane Abbott was the most abused female MP online during the 2017 General Election (Getty )

The charity is calling on the company to be more transparent about how it designs and implements the automated processed to identify online abuse against women, in the same way it has provided information about using its algorithms to combat misinformation around Covid-19. Amnesty says it will now be tracking Twitter’s progress via a Scorecard.

Michael Kleinman, director of the charity’s Silicon Valley Initiative, said: "It is totally in Twitter’s power to implement these changes that would make a real difference to millions of women’s experience on the platform. Twitter CEO Jack Dorsey needs to match words with action to show he is genuinely committed to making Twitter a safer place for women. We will continue to press the company until we see more changes that truly show that abuse against women is not welcome on the platform.”

In some ways, Twitter does agree with the charity. In response to Amnesty’s analysis, the company acknowledged it needed to do more. However, at the same time it says its combination of human moderation and tech allows it to take a more proactive response to online abuse.

In particular, Amnesty’s research has demonstrated how women from ethnic or religious minorities, marginalised castes, as well as LGBTQ women and non-binary individuals are disproportionality impacted by the abuse. An earlier study found that former shadow home secretary Diane Abbott was the most abused female MP online during the 2017 election, with one in 20 tweets featuring abuse.

Things haven't changed since then. Indian author and activist Meena Kandasamy said as a Tamil, mixed-caste women who speaks out against India’s caste system, she receives a “torrent of racist and misogynistic abuse, including rape threats.”

“Twitter always seems to be playing catchup and is too slow to address the different types of abuse women face. Twitter is a powerful place to express ourselves, but Twitter needs to do more to clean up the platform and make it a safe place for women," she says.

The company has other issues it needs to deal with too. This weekend, it was accused of using a “racist” image cropping algorithm that automatically focused on white faces over black ones. Several Twitter users tried it out on the platform, from posting images of US senator Mitch McConnell and former US president Barack Obama to see which image would be selected for the crop, to the Simpsons characters Lenny and Carl.

In a statement, a spokesperson for the company told the Guardian: “Our team did test for bias before shipping the model and did not find evidence of racial or gender bias in our testing. But it’s clear from these examples that we’ve got more analysis to do. We’ll continue to share what we learn, what actions we take, and will open source our analysis so others can review and replicate."

Read more

Women are receiving abusive tweets on average every 30 seconds