Junho Eum

Junho Eum

Hi! I’m a PhD candidate in Computer Science at George Washington University, advised by Adam Aviv. I work on usable security and privacy, with a focus on digital identity. I’m affiliated with NIST and have collaborated with researchers at MPI-SP, the University of Maryland, and CISPA.

What brought me to this field was a clip from Meeting You, a documentary in which a mother reunites in virtual reality with her 7 year old daughter, who had died three years earlier. The piece is controversial, but what struck me was the developers’ intent. They explained that they wanted to build something that could offer comfort to a grieving parent. This changed how I think about what technology is for.

Most technology is measured by how many people it reaches. I want to build technology where success means leaving fewer people behind.

My research keeps pointing to the same gap. Studying AI chatbots, I found the most frequent users trusted them most. Studying mobile driver’s licenses, I found users could share only the data a verifier needed but shared everything anyway because it was easier. In a recent survey of 853 US adults on online age verification, single sign-on was the strongest preference, again for convenience. People trade privacy for convenience until the trade exposes something they wanted to keep private.

We’ve signed up for so much that one more disclosure doesn’t feel like it matters. That’s the kind of problem I’m drawn to. The work below reflects my efforts at making the trade more visible to the people making it.

j.eum@gwu.edu CV Google Scholar GitHub LinkedIn

Publications

2026

User Perceptions of Online Age Verification Systems in the United States

Junho Eum, Jan Tolsdorf, Adryana Hutchinson, Smirity Kaushik, Adam J. Aviv
in review
2026

Identity in My Phone: An Interview Study on User Perceptions of Mobile Driver's Licenses

Junho Eum, Yee-Yin Choong, Arwa Alsahdi, Adam J. Aviv
in review
2026

Designing Against Over-Asking: Trust Signals, Friction, and the Limits of Interface Design in EU Digital Identity Wallets

Sumair Ijaz Hashmi, Junho Eum, Anna Calmbach, Stefan More, Jakob Heher, Katharina Krombholz, Matthias Fassl
in review
2026

How Probing for Problems and Bias Affects Perceptions of AI Chatbot Trustworthiness

Jan Tolsdorf, Alan F. Luo, Monica Kodwani, Junho Eum, Mahmood Sharif, Michelle L. Mazurek, Adam J. Aviv
ACM Conference on Fairness, Accountability, and Transparency · FAccT ’26 · accepted
acceptedFAccT

Investigates how the framing of survey questions about AI chatbots — specifically, whether participants are first prompted to consider potential problems or biases — affects their reported perceptions of trustworthiness. The work has methodological implications for how user studies on AI systems are designed and interpreted.

Full abstract available upon publication.

2025

Safety Perceptions of Generative AI Conversational Agents: Uncovering Perceptual Differences in Trust, Risk, and Fairness

Jan Tolsdorf, Alan F. Luo, Monica Kodwani, Junho Eum, Mahmood Sharif, Michelle L. Mazurek, Adam J. Aviv
Symposium on Usable Privacy and Security · SOUPS ’25
SOUPS

Public and academic discourse on the safety of conversational agents using generative AI, particularly chatbots, often centers on fairness, trust, and risk. However, there is limited insight into how users differentiate these perceptions and what factors shape them. To address this gap, we developed a survey instrument based on previous work. We conducted an exploratory study using factor analysis and latent class analysis on survey responses from n = 123 participants in the U.S. to offer an initial attempt at measuring and delineating the dimensionality of these safety perceptions.

Latent class analysis revealed three distinct user groups with sometimes counterintuitive perception patterns: The Hesitant Skeptics, The Cautious Trusters, and The Confident Adopters. We find that greater usage frequency of AI chatbots is associated with higher trust and fairness perceptions but lower perceived risk. Some demographic traits like sexual orientation, income, and ethnicity also had strong and significant effects on group membership. Our findings highlight the need for more refined measurement approaches and a more nuanced perspective on users’ AI safety perceptions regarding trust, fairness, and risk, particularly in capturing the kinds of experiences and interactions that lead users to develop their perceptions.

Read on USENIX
2025

On a Scale of 1 to 5, How Reliable Are AI User Studies? A Call for Developing Validated, Meaningful Scales and Metrics about User Perceptions of AI Systems

Jan Tolsdorf, Alan F. Luo, Monica Kodwani, Junho Eum, Mahmood Sharif, Michelle L. Mazurek, Adam J. Aviv
IEEE SPW Workshop on Technology and Consumer Protection · ConPro ’25
ConPro

Public discourse around trust, safety, and bias in AI systems intensifies, and as AI systems increasingly impact consumers’ daily lives, there is a growing need for empirical research to measure psychological constructs underlying the human-AI relationship. By reviewing literature, we identified a gap in the availability of validated instruments. Instead, researchers seem to adapt, reuse, or develop measures in an ad hoc manner without much systematic validation.

Through piloting different instruments, we identified limitations with this approach but also with existing validated instruments. To enable more robust and impactful research on user perceptions of AI systems, we advocate for a community-driven initiative to discuss, exchange, and develop validated, meaningful scales and metrics for human-centered AI research.

Read PDF (IEEE)