In 2020, the International Center for Research on Women (ICRW) and Quilt.AI formed a partnership and have spent the last several months researching online violence in India (here’s what we learned). With the onset of COVID-19, there has been an uptick in both online and offline violence. We took a more in-depth look specifically at online violence and behavior and compared what that looked like before and during the pandemic in India.
We recently sat down with the President of ICRW, Dr. Sarah Degnan Kambou and the CEO and co-founder of Quilt AI, Anurag Banerjee, to talk about how the partnership began. We also spoke with ICRW’s Senior Social and Behavioral Scientist, Dr. Laura Hinson; Quilt AI’s Senior Impact Director, Dr. Priyanka (Piya) Bhalla; and Quilt AI’s Program and Content Manager, Biva Rajbhandari, who have been working together on the project, digging deeper into online behaviors – specifically how people search for and post content about online violence in India.
Thank you all for joining us today. We’re looking forward to talking about the work your organizations have been doing together. But first, Sarah and Anurag, how did you two meet? And what was it about each other’s organizations that lent themselves to a partnership?
SARAH: Two years ago, Dr. Ravi Verma, ICRW’s Asia Director, and I attended a meeting in Jaipur focusing on creating positive social change among youth attending public middle schools. The Rajasthan State Government invited several organizations working on innovative approaches promoting gender equity to present at the meeting. It was my first exposure to Quilt.AI, a partner with ICRW on the Boys Stand Up Project. The presentation of the digital data was riveting, and the data themselves compelling. I dashed off an email to introduce myself to Anurag, and the rest is history.
ANURAG: We were thrilled to be working with ICRW on Boys Stand Up, which is a project focused on engaging boys in family planning in India. At Quilt.AI, we have worked extensively on studying gender equality and misogyny online. ICRW brought the same passion and experience to the issue with a rigorous research approach that I admired. We thought combining our expertise could pull up some interesting findings.
And why are partnerships like this one so important to you? What is the added value?
ANURAG: As more people access the Internet – especially girls – it is important to understand how we can create a safe online environment for them. To unleash the Internet for good can make powerful changes – from ensuring people are finding the right information on vaccines to services during a crisis.
SARAH: Through our on-going partnership with Quilt.AI, ICRW’s social scientists are exploring how to leverage the power of digital data. For several of our community studies, we are combining community-based mixed methods research with online data research. The synergies and complementarity are powerful, and they enhance our insight into community identity, concerns and needs, aspirations, forms of expression, and, in this time of crisis, local coping mechanisms.
Thank you, Anurag and Sarah. It sounds like a comprehensive approach to tackling online violence, a growing and pervasive issue. Laura, ICRW has been working on this for some time. Can you give us a little more context for the current work with Quilt.AI?
LAURA: I had been working for several years on a project related to what we call technology-facilitated gender-based violence (TFGBV), which is really the intersection of online and offline violence. In 2018, our team at ICRW took a deeper dive into researching what it is and how to measure it. And when we first started talking with Anurag and Piya, I thought it could be an interesting opportunity to continue that work through this partnership. Most of the work I do is survey work with individuals in person (and during COVID-19, over the phone). I wanted to know how we could learn about people’s experiences on a much larger scale, and Quilt.AI seemed to have the technology and skills to get us there.
How about your team, Piya?
PIYA: Similar to Laura, Quilt.AI has worked on understanding misogyny online and trends in physical, emotional, and sexual violence during the pandemic. We thought this study would complement that work because we had not looked at online violence specifically. Furthermore, ICRW’s expertise in TFGBV and the evidence-based framework helped structure our methodology in pulling digital insights.
What were you specifically trying to understand in this project, and how did you go about the work?
LAURA: We wanted to learn how to apply the technology-facilitated GBV conceptual framework to a new research question and ask about how people were searching for and posting about online violence, both before and during COVID. How might people shift their online communication and search habits during a pandemic context? We had already been reading reports that violence was increasing during COVID. And since violence online is still not fully recognized as the extension that it is of physical GBV, this was a chance to explore trends and bring more visibility to the issue.
BIVA: To understand search behavior, we first built a list of more than 6,000 keywords, such as ‘what is doxxing’. We categorized keywords under terms like ‘sexual harassment’ and measured how much each had increased or decreased in urban and rural areas. Finally, we pulled 8,000 posts related to online violence from Twitter and Facebook from before and during the pandemic and categorized them under specific types of violence (e.g. sharing image + private info without consent). We then quantified how much these categories had grown.
And what did you find, Biva?
BIVA: We found that online violence on Twitter tripled during the pandemic from 26% to 74%. Conversations around gender trolling and sexual harassment remained the largest at 47% and 35%, respectively. We also saw that, in both rural and urban areas, searches around harassment and exploitation were high, but doxxing was emerging quickly. For example, in Delhi, doxxing grew by 214% and in Mumbai by 211%. Finally, we found that, even when people are searching for help, they are met with limited local resources. Instead, they are redirected to foreign websites (e.g., British and American) or taken to local services like law firms and the Indian government cyber crime websites.
LAURA: In other words, there was a relatively large uptick during the pandemic of people searching for and posting about various forms of online violence, as compared to the months previous to the start of the pandemic. We are hoping that what we have found will start a broader conversation and open up opportunities for action to mitigate and prevent online violence.
That said, Anurag and Sarah, looking at online violence more broadly, what needs to happen to make inroads in this work? Given the complexities and rapid change of online platforms, how do we keep up and ensure that we disrupt violence in the digital space and ensure voices are not silenced?
ANURAG: We have to meet people where they increasingly are – online. Online interventions are critical in disrupting violence in the digital space and before they lead to offline consequences. The partnership between ICRW and Quilt.AI is an example of how we can make inroads. In particular, working with different organizations – research, tech, and government – we can scale up digital interventions that are centered around the voices of marginalized communities and build virtual safe spaces for them.
SARAH: Exactly. We need to engage a wide array of partners in this work – from governments, most of whom do not yet understand the complexity and consequences of this issue, to actors in the tech industry to women and girls, who experience online violence in new ways every day. Digital technologies can and should be safe for everyone. And we play a critical role in that by generating evidence like we have from this project and bringing people together to talk about concrete steps towards meaningful change.
Learn more about ICRW’s and Quilt.AI’s online violence project here.
And here are some other resources: