People of ACM - Meeyoung Cha

October 21, 2025

How is applying traditional computer science disciplines (e.g., network analysis, data mining) to the social sciences helping to make improvements in society?

Integrating classical computing methods into the social sciences allows entirely new ways to understand society. What excites me most is that this integration allows us to ask questions we simply couldn’t ask before. We have an abundance of data (such as internet traffic, social media posts, satellite imagery, and environmental data) which opens up astronomical opportunities for innovation, much like how computational biology once revolutionized the field by integrating diverse biological data.

When Twitter launched, I could immediately see an opportunity. Twitter (now X) soon became an experimental lab for studying social links and information flow for researchers. Our 2010 ICWSM paper, The Million Follower Fallacy, challenged assumptions about online influence and earned a Test of Time Award. It showed how large-scale data allows us to observe society in motion.

But beyond mere observation, we can gain new insights too. For instance, our analysis of rumor propagation on social media revealed the role of social links and led to important detection models combining structural, temporal, and linguistic signals. Computational social science, in this sense, offers a multi-dimensional lens for rethinking societal dy namics and designing evidence-based interventions.

What are the goals of the MPI-SP’s Data Science for Humanity group?

At the Data Science for Humanity Group, our goal is ambitious: to harness data science for meaningful social impact. We study how digital technologies shape human behavior, public discourse, and societal trust—particularly in areas such as misinformation, privacy, and algorithmic bias. We collaborate with non-governmental organizations (NGOs) to apply our findings to tangible, real-world scenarios. Our team also explores emerging challenges, including machine-induced harm, unlearning in transformer networks, and AI-generated persuasion.

To pursue frontier research, I try to bring together researchers from diverse fields such as geography, neuroscience, electronic engineering, journalism, human-computer interaction, economics, physics, and computer science. I look forward to seeing amazing, transformative insights emerge from these interdisciplinary collaborations.

In your most cited paper “I Tube, You Tube, Everybody Tubes: Analyzing the World's Largest User Generated Content Video System,” you, (along with co-authors Haewoon Kwak, Pablo Rodriguez, Yongyeol Ahn, and Sue Moon) provided an in-depth study of YouTube and other user-generated content systems. This paper was originally published in 2007 but received the ACM IMC Test of Time Award in 2022. Why have the insights from this paper remained so relevant?

Receiving the Test of Time Award was a deeply meaningful moment. It reminded us that research can have lasting relevance, especially when it helps illuminate technologies that change quickly but are built on enduring principles.

This was a paper I worked on as a PhD student. We studied the early dynamics of user-generated content platforms, primarily YouTube, which was just beginning to reshape how media was consumed. We examined patterns of user interaction, content distribution, and attention across both popular and long-tail niche videos.

What’s striking is how persistent the core challenges remain: algorithmic amplification, popularity bias, and the uneven visibility of content. While platforms have evolved over time, the fundamental questions around virality and fairness are still central to understanding today’s digital ecosystems.

One of your interests is combating fake news. What is a key challenge in this area? What emerging technologies will help with this effort?

One of the big challenges in combating fake news is its relentless evolution. Since our early research, misinformation has shifted across formats, languages, and platforms—and now, it’s increasingly AI-generated. The harm unfolds in real time, outpacing detection systems and leaving little room for intervention. That’s why our focus is on the deeper social and psychological forces that drive its spread.

As AI agents become more persuasive and autonomous, new questions arise about influence, trust, and human agency. These aren’t distant concerns—they are emerging risks we must address now, before they crystallize into systemic problems. Our recent work on detecting AI-generated text and assessing factuality is part of this effort to build foundations for more resilient information ecosystems. I hope to see future research that more deeply reflects the ethical, cultural, and cognitive dimensions of misinformation.

As a mentor, what is an important piece of advice you offer younger colleagues just starting out in the field?

One piece of advice I often share is this: give yourself a million chances. In the early stages of a career, it’s easy to think, “I don’t have this skill, so I’ll never get that job.” But I encourage students to flip that mindset. What if you gave yourself five years? Could you learn, grow, and reach that goal? What about three years? Even two?

Once you allow for that possibility, everything begins to shift. You start to see the future not as a fixed outcome, but as something you can shape. Then, ask yourself “What would I need to do six months before that moment or one year before?” With that perspective, you realize that most things are possible; it’s often just a matter of time. And because it’s a long journey, be kind to yourself along the way!