Office hours with… Daniel Karell

Yale sociologist Daniel Karell seeks to understand how online behavior generates real-world harm. In studying the nature and culture of extremist online communities, he has explored the internet’s role in fomenting radicalism in Afghanistan and the United States.

We caught up with him for the latest edition of Office Hours, a Q&A series that introduces newcomers to the Yale faculty to the broader university community.

How do you describe your research interests to people outside of academia?

I try to make my research relatable by emphasizing how my current work helps uncover potential offline harms of online behavior. Specifically, I look at how online content or communities can trigger physical violence and other types of harm in the real world.

I currently live near Seattle and I interact a lot with people who work in the tech industry. They’re often interested in another dimension to my research, which involves using computational and big-data approaches to measure culture, especially how people express themselves culturally online.

How did you become interested in the relationship between online behavior and political extremism?

I’ve been working for many years on extremism and radicalism in Afghanistan and Pakistan. I drew a lot on print media because several of the radical groups, including the Taliban, originate in the 1980s and 1990s. As that research began entering its late stages, I saw a lot of parallels between what I was reading in that print media from decades ago and what’s appearing in the digital media produced by extremists today.

At the same time, there was a big-data and computational revolution happening in the social sciences. I became interested in the technical methodology and tools available for examining the effects of digital media on violence and extremism.

What does the future hold for Afghanistan now that the Taliban has retaken control of the country?

Despite the Taliban's recent assurances, I think it’s likely that its rule will be similar to the previous incarnation. If anything, the scholarship on radicalization leads us to expect that the Taliban would be more extreme after a longer period of conflict and its opponents’ embrace of drone warfare, indefinite detention, and torture. Yet, the world and Afghanistan have changed a lot since 2001. The country is far more connected to the rest of the world, which is itself more interconnected, than it was in the 1990s.

One of the great successes in Afghanistan over the last 20 years has been the emergence of a lively and robust telecommunications and media sector. In addition, many Afghans are online and use social media. As a result, Afghans are going to be constantly aware of ideas and policies that run contrary to the Taliban's, as well as alternative voices and ways of life from around the world. We’ll see how the Taliban navigates these threats to its hegemony, traditionalist ideology, and form of rule. 

What’s something you’re currently working on?

A lot of people believe that online behavior has significant offline consequences. For example, people might assume that the January 6 riot at the U.S. Capitol was organized online. Whether or not this is true, it’s actually very difficult to match social media activity to offline events.

Parler is a rightwing micro-blogging platform that was de-platformed by Amazon Web Services in January after the riot. Before the de-platforming, internet archivists scraped Parler for data, creating a geo-referenced record of all activity that had occurred on it. I’ve been working on a paper with some geographers to plot all this activity and see if it corresponded with right-wing protests that occurred across the United States during 2020 and 2021. We find strong evidence that Parler activity increased subsequent right-wing protests.

You taught a course last semester for the YData program on measuring culture. How did it go?

It was the first time I’ve taught the course, so there were some wrinkles to iron out, but the students seemed very interested in both the methodological side of working with digital data and the substantive side of using data to answer interesting questions about culture. 

By far, students were most interested in data from TikTok, the video-sharing platform. I knew very little about TikTok, but my amazing teaching fellow, Jeffrey Sachs, figured out how to pull data from the app.

Jeff and I actually wrote a paper with one of the students, Rahshemah Wise, that examines whether people have built positive communities on TikTok during the pandemic in which they feel comfortable expressing their true selves, such as by sharing “coming out” videos about their gender identity or sexual orientation.  As part of the project, we made available a dataset of 4.8 million TikTok videos. As far as we know, it’s the largest TikTok dataset that exists.

Amid the challenges of the pandemic, have there been there any pleasant surprises for you in the last year, professionally or personally?

My previous job was at New York University Abu Dhabi. I made a lot of international connections in Abu Dhabi and I have projects with collaborators based in the Middle East, Europe, and across the United States. The fact that video meetings have become the norm has really accelerated my projects with these international collaborators, which is great.

Personally, I’ve enjoyed the opportunity to spend more time with my kids. I have a 4-year-old daughter and a 2-year-old son. We have been living in Washington next to the North Cascades National Park since the pandemic began. We’ve spent a lot of time outdoors. This means a lot of backcountry skiing, rock climbing, and mountain biking. I taught my kids how to ski this winter, which was a lot of fun. My 4-year-old is already a pretty fast mountain biker.