6/24/2025 Bruce Adams
CS professor Eshwar Chandrasekharan is a recipient of an NSF CAREER grant for his work developing new methods to support community-determined behavior norms in online communities consistent with the goal of helping communities reward behaviors that serve the community's goals.
Written by Bruce Adams
Which works better when moderating online behavior: the carrot or the stick?
CS professor Eshwar Chandrasekharan is a recipient of an NSF CAREER grant for his work developing new methods to support community-determined behavior norms in online communities consistent with the goal of helping communities reward behaviors that serve the community's goals. As he describes, “my students and I are exploring new ways to foster welcoming online spaces, and we're excited to advance this research with the support of the award.
In describing his work, Chandrasekharan says that
“After a decade of researching online moderation, I see a natural next step in exploring how positive reinforcement can proactively encourage the kinds of behaviors communities value. Rather than focusing solely on punishing undesirable content, this approach aims to foster healthier, more supportive online spaces.
Punitive moderation strategies can improve discourse, but they fall short in two key ways. First, because they intervene only after harm has occurred, harmful content often remains visible before it’s addressed. More importantly, these approaches don’t actively cultivate a positive, prosocial environment—one that encourages meaningful, respectful engagement and supports the long-term health of online communities.”
Chandrasekharan’s research matches computer science with insight from the social sciences. “This approach builds on well-established principles in behavioral psychology: positive feedback has long been used to shape behavior. When applied to online spaces, it offers a more transparent and constructive way to foster engagement—one that can reduce the burden on moderators by shifting the focus toward promoting what’s good, rather than only reacting to what’s harmful.”
Upvotes, badges, or highlighted comments can all serve to publicly highlight prosocial behavior in a community and clearly display community norms.
The team aims to create a public application programming interface that enables developers and moderators to integrate the computational approaches developed in this research into their own communities. Easy access to pre-trained models and carefully designed, human-centered tools will enable moderators to discover valuable, normative behavior within their communities and respond with timely, positive reinforcement.
Grainger Engineering Affiliations
Eshwar Chandrasekharan is an Illinois Grainger Engineering professor of computer science.