Postdoctoral Associate
Institute for Artificial Intelligence and Data Science
Department of CS & Engineering
University at Buffalo, SUNY
qingxiao[at]buffalo[dot]edu
Buffalo, NY
Hello, beautiful people!
I’m Qingxiao, and I believe technology should amplify human potential, not replace it. As a researcher, designer, and innovator specializing in human-centered AI, I‘m committed to pioneering responsible AI systems that democratize emerging technologies and drive meaningful societal impact.
My research interests are:
My background spans academia and industry: I hold a PhD in Information Sciences at the University of Illinois Urbana-Champaign (UIUC), served as Director of Data Science at a B2B AI company, and received academic training in social science theories and methods at the Chinese University of Hong Kong (CUHK).
“See, Feel, Think, Act” is my simple formula for navigating life’s journey. I’m excited to collaborate on research that bridges theory and practice, centers human needs, and creates AI systems that truly serve communities.
LATEST NEWS
Publications: [Google Scholar]
Evaluating UX in the context of AI’s complexity, unpredictability, and generative nature presents unique challenges. How can we support HCI researchers to create comprehensive UX evaluation plans? In this paper, we introduce EvAlignUX, a system powered by large language models and grounded in scientific literature, designed to help HCI researchers explore evaluation metrics and their relationship to research outcomes. In a world where experience defines impact, we discuss the importance of shifting UX evaluation from a “method-centric” to a “mindset-centric” approach as the key to meaningful and lasting design evaluation.
This paper asks a core design and evaluation question: when and how should AI engage in human interaction, and how do dyadic versus polyadic configurations shape user experience? As conversational agents move from one-on-one interactions to mediating multi-party social dynamics, they introduce distinct UX challenges around coordination, social roles, and interpretability. We synthesize UX-focused ACM studies to establish a design-oriented understanding of how different interaction configurations affect human experience and inform future human-AI interaction system evaluation approach.
Participatory Design (PD) aims to empower users by involving them in various design decisions. However, it was found that the PD’s evaluation criteria are usually set by the product team and used only at the end of a design process without adequate user participation. To address this issue, we proposed introducing UX evaluation metrics into design materials at the participatory design INPUT phase.
This artistic collaborative project introduces a tangible design fiction in which participants take part in a speculative 2052 dining experience that involves consuming biohybrid flying robots, using multisensory performance and ritual to probe how people reason about future food technologies. The study shows that embodied, ambiguous encounters surface nuanced ethical, cultural, and affective negotiations around edibility, sentience, and sustainability, advancing HCI methods for evaluating speculative AI–biohybrid systems beyond abstract discussion.
In speech-language service and care contexts, my research examines how AI systems can be responsibly designed and evaluated to support speech-language pathologists, caregivers, and children across clinical, home, and training settings. Rather than optimizing for automation or performance alone, this program of work investigates how expertise-aligned AI shapes trust, confidence, emotional engagement, and professional judgment in speech-language services. This research is primarily supported by the NSF National AI Institute for Exceptional Education and has most recently received a 2025 research grant from the Organization for Autism Research (Co-PI).
In the law enforcement context, this collaborative project examines what is gained and lost when high-stakes, trauma-informed communication training shifts from live, actor-based role-play to AI-based simulations. Through a mixed-methods study with police recruits, it shows that AI is most effective when strategically sequenced with human simulations, functioning as a complementary scaffold that reshapes emotional engagement, preparedness, and learning rather than replacing human realism.
In the medical training context, this collaborative project investigates when an AI facilitator should proactively intervene versus remain user-initiated in immersive medical training, through a mixed-methods study comparing proactive and on-demand AI guidance in an XR lumbar puncture simulator. While learning outcomes were similar, qualitative findings reveal that the perceived value of AI proactivity depends on task phase, cognitive load, and learner preference, leading to a boundary framework for calibrating AI intervention in high-stakes, cognitively demanding training environments.
In the law enforcement context, we present EMSIM, an AI-driven VR system for de-escalation training that integrates an LLM-based, rubric-guided empathy evaluator to dynamically shape scenario progression and provide real-time, feedback-rich reflection for both trainees and instructors. Through mixed-methods evaluation with law enforcement trainees and trainers, the study demonstrates how AI-enabled XR systems can support exploratory learning and system-level evaluation of communication strategies, while surfacing key design tradeoffs around controllability, feedback timing, and realism in high-stakes human–AI training systems.
Best Paper Award 🏆
This empirical study serves as a primer for interested service providers to determine if and how Large Language Model (LLM) technology will be integrated for their practitioners and the broader community. The insights pave the way for synergistic and ethical human-AI co-creation in service contexts.
This study examines the impact of AI-generated digital clones with self-images (AI-clones) on enhancing perceptions and skills in online presentations. A mixed-design experiment with 44 international students compared self-recording videos (self-recording group) to AI-clone videos (AI-clone group) for online English presentation practice.
RQ: How do people (e.g., victims, attackers, bystanders, or spectators) respond to safety risks posed by virtual avatars, and what are the design implications for avatar-based human-human interactions?
Understanding emerging safety risks in nuanced social VR spaces and how existing safety features are used is crucial for the future development of safe and inclusive 3D social worlds. Prior research on safety risks in social VR is mainly based on interview or survey data about social VR users’ experiences and opinions, which lacks “in-situ observations” of how individuals react to these risks. Using two empirical studies, this paper seeks to understand safety risks and safety design in social VR. In Study 1, we investigated 212 YouTube videos and their transcripts that document social VR users’ immediate experiences of safety risks as victims, attackers, or bystanders. We also analyzed spectators’ reactions to these risks shown in comments to the videos. In Study 2, we summarized 13 safety features across various social VR platforms and mapped how each existing safety feature in social VR can mitigate the risks identified in Study 1. Based on the uniqueness of social VR interaction dynamics and users’ multi-modal simulated reactions, we call for further rethinking and reapproaching safety designs for future social VR environments and propose potential design implications for future safety protection mechanisms in social VR.
RQ: How can we design a chatbot to mediate emotional communication?
Many couples experience long-distance relationships (LDRs), and “couple technologies” have been designed to influence certain relational practices or maintain them in challenging situations. Chatbots show great potential in mediating people’s interactions. However, little is known about whether and how chatbots can be desirable and effective for mediating LDRs. In this paper, we conducted a two-phase study to design and evaluate a chatbot, PocketBot, that aims to provide effective interventions for LDRs. In Phase I, we adopted an iterative design process by conducting need-finding interviews to formulate design ideas and piloted the implemented PocketBot with 11 participants. In Phase II, we evaluated PocketBot with eighteen participants (nine LDR couples) in a week-long field trial followed by exit interviews, which yielded empirical understandings of the feasibility, effectiveness, and potential pitfalls of using PocketBot. First, a knock-on-the door feature allowed couples to know when to resume an interaction after evading a conflict; this feature was preferred by certain participants (e.g., participants with stoic personalities). Second, a humor feature was introduced to spice up couples’ conversations. This feature was favored by all participants, although some couples’ perceptions of the feature varied due to their different cultural or language backgrounds. Third, a deep talk feature enabled couples at different relational stages to conduct opportunistic conversations about sensitive topics for exploring unknowns about each other, which resulted in surprising discoveries between couples who have been in relationships for years. Our findings provide inspirations for future conversational-based couple technologies that support emotional communication.
Fun Facts
Share writing time with my wonderful dog, who appears to be ever-curious.
When not experimenting on AI, I’m experimenting with new recipes in the kitchen.
My workspace is clutter-free.
My dog’s AI Clone.
🐾 You’ve Reached The Tail End 🐶
See you next time 🙂