The Good Bots: One Cyborg That’s Killing the Trolls With Kindness

Social media represents a powerful opportunity to increase the visibility of women in politics—but this should not come at any cost. By educating, validating and posting positivity, ParityBOT hopes to challenge the online culture of toxicity, one tweet at a time.

parityBOT, 16 days of activism, online harassment, women journalists, women, twitter
(Pixy / Creative Commons)

When the news of President Trump’s positive COVID diagnosis broke this fall, Twitter posted an important clarification for user behavior:

“Tweets that wish or hope for death, serious bodily harm or fatal disease against *anyone* are not allowed and will need to be removed.”

The hypocrisy of this statement was soon pointed out by women everywhere, considering how often death threats are directed at women in politics, journalism and ultimately any profession or hobby where they dare to have and share an opinion online. 

As Congresswoman Alexandria Ocasio-Cortez tweeted, “So … you mean to tell us you could’ve done this the whole time?”

As much as social media platforms present themselves as objective facilitators for the spread of ideas and communication, anyone who has spent time on one knows that the rules of engagement are anything but unbiased. Which messages get amplified and which are met with derision—or worse—is often more a response to the messenger than the message itself.

Online Abuse is Rampant—and it Translates Offline, Too

Toxicity on the internet is a problem for everyone—but marginalized communities are especially negatively impacted by it.

While women receive more abuse online, it also tends to be more violent in nature. A 2018 study from Amnesty International called Troll Patrol showed that 7.1 percent of tweets sent to the women in the study were “problematic” or “abusive.”

Women of color were 34 percent more likely to be mentioned in abusive or problematic tweets than white women. Black women were disproportionately targeted, being 84 percent more likely than white women to be mentioned in abusive or problematic tweets.

This is a salient issue when tackling gender disparity in politics, where an online presence can be critical for fundraising, reaching voters and engaging with constituents. Women are globally underrepresented in politics, despite evidence that greater gender parity improves outcomes in important indicators, such as health and childcare.

While efforts such as the UN’s Beijing Declaration and Platform for Action aim to achieve gender balance in leadership roles, facing disproportionate harassment and abuse online is an additional hurdle to clear. Women are left out—or pushed out—of these online conversations when they must wade through constant vitriol and bad-faith discourse. Research from the U.S., U.K. and Australia has found that women politicians receive far more directed abuse and incivility online than their male peers.

Bots Can Validate the Toxicity Experienced by Women

The desire to encourage more women to run for office led to the development of ParityBOT, a Twitter bot created by Areto Labs to counteract online abuse by sending human-written positive tweets. Using a machine learning model to automatically evaluate a tweet’s sentiment, ParityBOT scores tweets for toxicity and responds by sending a positive tweet if a certain threshold of negativity is breached. The positive tweets are contributed by humans and vetted by the Areto Labs team, and offer supportive, encouraging messages for women candidates.  

parityBOT, 16 days of activism, online harassment, women journalists, women, twitter
ParityBOT. (Twitter)

The idea behind ParityBOT is that by countering the hostility in social media with positivity, it contributes to disrupting the negative feedback loop.

Currently, there is little research on the effectiveness of interventions in cases of online abuse against women leaders. To this end, the team behind ParityBOT interviewed women candidates after its initial run in the 2019 Canadian elections. Support for the ParityBOT concept was overwhelmingly positive, particularly as a way to visibly encourage more women to run for office.


Here at Ms., our team is continuing to report through this global health crisis—doing what we can to keep you informed and up-to-date on some of the most underreported issues of this pandemic. We ask that you consider supporting our work to bring you substantive, unique reporting—we can’t do it without you. Support our independent reporting and truth-telling for as little as $5 per month.


Thousands of Toxic Tweets Were Sent to Women in Politics in October

After its positive reception in Canada, Areto Labs launched ParityBOT_NZ for the 2020 New Zealand federal election, followed by ParityBOT_US.

From Sept. 29 to Nov. 17 this year, ParityBOT_US tracked over 12 million tweets sent to 376 women candidates in the U.S. election. Toxic tweets were identified by ParityBOT’s sentiment analysis model, and were also categorized by type of abusive content, such as sexually explicit or threatening. Of the monitored tweets, more than 12,000 were flagged as toxic with an extremely high level of certainty, triggering a positive response from ParityBOT_US. 

High profile candidates received far more attention and negative tweets than lesser known ones. As a woman of color, vice presidential nominee Kamala Harris was already more likely to receive online abuse, and this was undoubtedly compounded by her historic running as the first woman vice presidential nominee.

parityBOT, 16 days of activism, online harassment, women journalists, women, twitter
During the week of the vice presidential debate, the number of abusive tweets sent to Harris were higher than average, with a significant spike the day after the vice presidential debate, and again on Nov. 8, the day after the election was called for her and Joe Biden. (Gage Skidmore / Flickr)

When Harris was chosen as Joe Biden’s running partner, racial and gender-motivated attacks directed on her at social media escalated immediately, with experts believing they would only increase throughout the campaign.

The evidence from ParityBOT_US indicates these concerns were, in fact, well-founded: During the week of the vice presidential debate, the number of abusive tweets sent to Harris were higher than average, with a significant spike the day after the vice presidential debate, and again on Nov. 8, the day after the election was called for her and Joe Biden.

In fact, the volume of abusive tweets sent to women candidates increased substantially that week, and was even larger the following week. Toxic tweets weren’t limited to the campaign period, and appear to have actually ramped up after the election.

Evidence that shows the negativity continues, or even increases once in power, can be another daunting reason why women are dissuaded from participating in public office—particularly those in more senior or high profile roles.

It also underlines the continual need to monitor these interactions and support women once they are in office—not only to show them support, but also to show women with future ambitions for public office that someone is watching, someone has their back, and politics is an important space for women like them, from diverse backgrounds, to participate in. 

Challenging the Culture of Toxicity for Sustainable Improvement Online

Social media platforms themselves are complicit in gaslighting women about the nature of their online interactions, failing to take seriously the abuse that ParityBOT, among many others, have documented. These platforms are not simply neutral facilitators of communications, but rather shape the dynamic of these conversations with their policies. By creating a record of these negative interactions, ParityBOT supports women in validating their online experiences while showing they aren’t alone.

Social media represents a powerful opportunity to increase the visibility of women in politics and magnify their platforms, but this should not come at any cost. By educating, validating, and posting positivity, ParityBOT hopes to challenge the online culture of toxicity, one tweet at a time.

ParityBOT was created by Areto Labs, a women-led social tech enterprise. We build bots that post positive messaging and behavioural nudges in digital communities like Slack and Twitter, and at the same time, analyze sentiment to provide actionable insights to help businesses build better brands and cultures. Get in touch to learn more.

Get caught up on the Ms. 16 Days of Activism collection.

You may also like:


The coronavirus pandemic and the response by federal, state and local authorities is fast-movingDuring this time, Ms. is keeping a focus on aspects of the crisis—especially as it impacts women and their families—often not reported by mainstream media. If you found this article helpful, please consider supporting our independent reporting and truth-telling for as little as $5 per month.

About

Sam Work is an AI analyst and researcher, and is currently Head of Research at Areto Labs. She has an MSc from University College London.