Chris Bail studies Polarization in the emerging field of computational social science (CSS). Combining the epistemes of social science with the technologies of computer science, the method of CSS recognizes the complex systems involved in big social issues and attempts to make social research easier and more efficient, by drawing on computer science to process data as well as simulate and analyze social behaviors.

Tackling Expanding Gaps

Chris is the founder of the Polarization lab at Duke University, a collaborative research initiative that brings experienced researchers together from social sciences, computer science and statistics to tackle the expanding gaps that threaten the integrity of our social structures. 'Polarization' is a global phenomenon, but is often discussed in an American context, as their bi-partisan political structure reproduces polarization in the more explicit, literal sense.

‘I would say the US is too often the focus of research on this topic. My interest in polarization is related to the study of social media and its impact on behaviors and attitudes’.

Highlighting research in the US and Europe on the global impact of social media, Bail notes that early evidence indicates significant regional variations, making it a crucial topic for exploration. ‘This shows that if we took the US as the only plane of reference, we would end up with a very distorted view of how the effects of social media and polarization unfold’.

Beyond the US

Growing up in The French Congo, China and Switzerland, Chris has first-hand experience with sociopolitical realities outside of the US. Moreover, coming from an intellectual background in cultural sociology, his research interests remain broad. ‘My original ambition in graduate school was to study culture of immigration and development in west Africa. Somehow I wound up studying polarization in the US. But before I became a professor I was doing work on anti-muslim extremists in the US and UK. So even though my recent research is heavily focused on the US, in general I have a strong appreciation for the need to take things beyond the US’, he reassures.

Team Science

As a computational social scientist, Chris is committed to the virtue of collaboration through the model of Team Science. He reminds us that while it is commonplace in the natural sciences to conduct research in groups - with several co-authors attached to a publication - research in the social sciences and humanities has generally been done with sole authors. According to Chris, this is detrimental rather than beneficial.

‘If we look at science at large, tens of thousands of studies across many different disciplines show that when a team of researchers are involved, and they have a diverse background, they produce more pathbreaking research, receive more citation and they get more media attention. It’s intuitive that this allows us to fill in the gaps; if we see science as a big network of ideas, then someone needs to pull the ideas together.’

Diversity Breeds Better Science

An example of  team science is seen in the Summer Institutes of Computational Social Science, a global training program Chris co-founded in 2017. It unites young scholars from diverse academic backgrounds to collaborate on solving various issues through computational social science tools. ‘One of the reasons I wanted to do that, was because I read a lot of research about how diversity breeds better science. All kinds of diversity: interdisciplinary diversity, gender diversity, racial diversity, geographic diversity.

Interdisciplinarity has become a bit of a buzzword, used frequently in the abstract in close to all academic circles – some might say this puts it at risk of losing its meaning. ‘What I’m describing is a way to take the underlying principles of interdisciplinarity and apply it in a practical and generative way.’

Research Speed Dating

In these Summer Institutes, this becomes more explicit. ‘We have the participants of the summer institute come together and do something called research speed dating’.

Initially, participants spend one week on coursework. In the second week, they put their research interests into a Google sheet, and an algorithm forms small groups, aiming for either maximum similarity or maximum dissimilarity. Over the next three hours, these groups generate research questions every 20 minutes, resulting in a selection of viable research teams for participants to join and continue their work.

‘The idea is to encourage interdisciplinarity by formally structuring the way interdisciplinary collaboration happens. We can encourage people to confront their similarities and their differences, and out of this process we see that people develop really fascinating interdisciplinary research agendas.’

Understanding Social Media

Bail has a lot to share, when it comes to his view on the past, present and future of computational social science. ‘I used to have to convince people to care about social media. When I would give lectures, audience members would ask ‘Why should we care about social media? It’s just people talking online.’ People would say it wasn’t serious, it wasn’t meaningful.

‘Now I ask my students: can you imagine studying politics without social media? Can you imagine studying most things without understanding social media? And that’s been a 10 year change! That’s a remarkable shift. Usually science doesn’t change that quickly.’

CSS and Artificial Intelligence

In the present, there are huge challenges around access to data, especially when it comes to generative Artificial Intelligence (AI). ‘Increasingly, CSS is going to be involved in conversations about what generative AI is, what are its limitations? I think it’s going to be a part of the future of artificial intelligence as well, especially if AI moves towards general intelligence.

If that’s the goal – and I’m not at all sure it should be - I can’t imagine AI progressing without incorporating some model of society and human behavior. It’s one thing for a generative AI (such as ChatGPT) to regurgitate some information from the social science, but that’s very different than applying it to understand, in a dynamic and emergent way, how to navigate a social situation. For now, it’s about figuring out how CSS can help protect the world against generative AI, but also shape the future of AI’.