Alicia (Hyun Jin) Lee

Alicia (Hyun Jin) Lee

Auto-generated profile, based on publicly available information

Alicia (Hyun Jin) Lee

Human-Computer Interaction Institute, Carnegie Mellon University, Pittsburgh, Pennsylvania, USA

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 July 2024
Verified
Conference Paper
Open access
MCMai Lee Chang
Mai Lee Chang
ALAlicia (Hyun Jin) Lee
Alicia (Hyun Jin) Lee
NHNara Han
Nara Han
AHAnna Huang
Anna Huang
HSHugo Simão
Hugo Simão
ASAaron Steinfeld
Aaron Steinfeld
The population of older adults experiencing cognitive decline is growing faster than the number of workers who can care for them. Artificially intelligent (AI) agents could assist these older adults, keeping them in their homes longer. For this to happen, older adults must be willing to adopt and rely on agents. Would they trust an agent that might need to report their decline to others? We conducted a speed dating study exploring the impact of agent affiliation (i.e., who the agent should work for). Our healthy and declining participants reacted positively to the idea of agents supporting them. They particularly recognized how the agent would reduce the burden placed on their family caregivers. They viewed affiliation to be dynamic, shifting from the declining older adult and orienting more to their caregivers over the course of cognitive decline. They envisioned the agent modifying its decision-making process to be like their caregivers’.

2 May 2025 in Proceedings of the ACM on Human-Computer Interaction
Verified
Journal Article
Open access
MCMai Lee Chang
Mai Lee Chang
SRSamantha Reig
Samantha Reig
ALAlicia (Hyun Jin) Lee
Alicia (Hyun Jin) Lee
AHAnna Huang
Anna Huang
HSHugo Simão
Hugo Simão
ASAaron Steinfeld
Aaron Steinfeld
As the population of older adults increases, there is a growing need for support for them to age in place. This is exacerbated by the growing number of individuals struggling with cognitive decline and shrinking number of youth who provide care for them. Artificially intelligent agents could provide cognitive support to older adults experiencing memory problems, and they could help informal caregivers with coordination tasks. To better understand this possible future, we conducted a speed dating with storyboards study to reveal invisible social boundaries that might keep older adults and their caregivers from accepting and using agents. We found that healthy older adults worry that accepting agents into their homes might increase their chances of developing dementia. At the same time, they want immediate access to agents that know them well if they should experience cognitive decline. Older adults in the early stages of cognitive decline expressed a desire for agents that can ease the burden they saw themselves becoming for their caregivers. They also speculated that an agent who really knew them well might be an effective advocate for their needs when they were less able to advocate for themselves. That is, the agent may need to transition from being unremarkable to remarkable. Based on these findings, we present design opportunities and considerations for agents and articulate directions of future research.

13 October 2025
Verified
Conference Paper
Open access
ALAlicia (Hyun Jin) Lee
Alicia (Hyun Jin) Lee
CIChaco Iwase
Chaco Iwase
MXMaria Xu
Maria Xu
CHChengguizi Han
Chengguizi Han
JHJessica Hammer
Jessica Hammer

13 April 2026
Verified
Conference Paper
Open access
QWQiaosi Wang
Qiaosi Wang
JKJini Kim
Jini Kim
ASAvanita Sharma
Avanita Sharma
ALAlicia (Hyun Jin) Lee
Alicia (Hyun Jin) Lee
JFJodi Forlizzi
Jodi Forlizzi
HSHong Shen
Hong Shen
Theory of Mind (ToM)—the ability to infer what others are thinking (e.g., intentions) from observable cues—is traditionally considered fundamental to human social interactions. This has sparked growing efforts in building and benchmarking AI’s ToM capability, yet little is known about how such capability could translate into the design and experience of everyday user-facing AI products and services. We conducted 13 co-design sessions with 26 U.S.-based AI practitioners to envision, reflect, and distill design recommendations for ToM-enabled everyday AI products and services that are both future-looking and grounded in the realities of AI design and development practices. Analysis revealed three interrelated design recommendations: ToM-enabled AI should 1) be situated in the social context that shape users’ mental states, 2) be responsive to the dynamic nature of mental states, and 3) be attuned to subjective individual differences. We surface design tensions within each recommendation that reveal a broader gap between practitioners’ envisioned futures of ToM-enabled AI and the realities of current AI design and development practices. These findings point toward the need to move beyond static, inference-driven approach to ToM and toward designing ToM as a pervasive capability that supports continuous human-AI interaction loops.

28 August 2026 in JMIR Preprints
Verified
Preprint
Open access
MSMark J Seidman
Mark J Seidman
MBMalaz R Boustani
Malaz R Boustani
EREdmond Ramly
Edmond Ramly
RHRichard J Holden
Richard J Holden
Melissa PangelinanMelissa Pangelinan
Melissa Pangelinan
YSYiyu Shi
Yiyu Shi
UNSTRUCTURED Digital therapeutics are digital technology-based interventions designed to treat, manage, or prevent medical or behavioral conditions. Early products delivered fixed educational, behavioral, or monitoring content. Newer products increasingly resemble adaptive software ecosystems that incorporate large language models, multimodal real-time data, memory, retrieval, tool use, and multi-agent coordination. Such systems change in two distinct ways. The first is planned: a team updates the model, data, configuration, or permissions. Regulatory mechanisms such as predetermined change control plans govern these updates. The second is not planned. An agentic system can act differently at runtime with nothing newly deployed, because retrieval surfaces different context, memory accumulates, or agents select different tools. These changes cannot be enumerated in advance, and neither clinical stage models nor software governance frameworks specify how to develop, evaluate, or retire products that exhibit unplanned behavior. We propose an Agile Stage Model for minimally viable development and governance of digital therapeutics, including those built on agentic artificial intelligence. The model draws on the NIH Stage Model, software-as-a-medical device principles, Agile Science and its derivatives, the artificial intelligence agent literature, intelligent delegation, and recent work on agentic artificial intelligence for executing clinical trials. It extends the Agile Analytic Engine, an architecture that connects population evidence to a local treatment-to-target decision contract specifying target, baseline, action, reassessment interval, and response rule, and then to a governed chain of minimally viable experiments. The model comprises six stages: minimum viable specification (Stage 0), minimum viable product and internal validation (Stage 1), first-in-human sensor-action testing (Stage 2), local real-world minimally viable experiment (Stage 3), population effectiveness and cross-site learning (Stage 4), and lifecycle stewardship and deimplementation (Stage 5). The stages are learning gates rather than a fixed sequence, and teams may advance, repeat, revise, pause, or stop. At each stage the team asks a single question: what is the smallest valid experiment that can inform the next decision? For agentic systems, every stage adds explicit requirements for bounded authority, human oversight, permissioning, verification, and monitoring that separates degradation in performance from change in system behavior. One rule governs the model: standardize only what protects patient safety and the integrity of evidence and let everything else adapt to context. Oversight should scale with the severity, irreversibility, and reach of potential harm rather than being uniform across products. The model is conceptual and requires prospective evaluation.

24 October 2026
Verified
Conference Paper
Open access
HSHugo Simão
Hugo Simão
BNBeatriz Nogueira
Beatriz Nogueira
ALAlicia (Hyun Jin) Lee
Alicia (Hyun Jin) Lee
Alexandre BernardinoAlexandre Bernardino
Alexandre Bernardino
JFJodi Forlizzi
Jodi Forlizzi
Tiago GuerreiroTiago Guerreiro
Tiago Guerreiro