BVI Labs
PUBLICATIONS & PROJECTS

생각을 검증하고,
지식을 나눕니다.

BVI Labs의 논문과 프로젝트 성과를 주제와 연도별로 살펴보세요.

02
01
PAPER · Short-form video, perceived agency, swiping control, consistency, mediation2026

Swiping Control and Thematic Consistency in Short-Form Video Platforms: Implications for Perceived Agency and User Experience

김태영

Short-form video platforms have transformed digital viewing into an interactive, gesture-based experience, yet the psychological effects of their design remain unclear. This study examined the effects of interface control (manual swiping vs. automatic advancement) and thematic consistency (consistent vs. inconsistent content) on user experience. In a 2 × 2 between-subjects experiment, participants interacted with a short-form video feed and reported perceptions of agency, satisfaction, expectation congruency, and relevance. Results showed that manual swiping increased perceived agency, and this effect was strongest when videos followed a coherent thematic flow. Thematic consistency also increased expectation congruency, relevance, and satisfaction. These findings suggest that users’ experiences of agency depend not only on having the opportunity to act but also on whether system responses are experienced as coherent and predictable. This study highlights the importance of understanding how users experience the relationship between their actions and system responses during continuous interaction on short-form video platforms

International Journal of Human–Computer Interaction
02
PAPER · Recommendation, agency perception, perceived assistance, alternative options, reasoning2025

Strategizing AI Recommendations: Focusing on the Role of Agency and Assistance Perception in Enhancing Engagement

김태영

Artificial intelligence (AI) is central to digital platforms, particularly in personalized recommendation systems to enhance user interactions. This study investigates how AI recommendation strategies—presenting alternatives and providing reasoning—impact user responses, considering the moderating role of agency perception (control over one’s actions) and the mediating role of perceived assistance (the sense of being supported by AI). An online experiment manipulated the presence of alternatives and reasoning to measure perceived agency and assistance. Results show that presenting alternatives had a direct negative impact on users with high agency, viewing alternatives as intrusive or misaligned. In contrast, users with low agency experienced positive outcomes, perceiving greater assistance and showing improved user engagement and satisfaction. These findings reveal a dual-pathway mechanism, where the same AI strategy elicits divergent responses based on agency perception, emphasizing the need for personalized recommendation designs

International Journal of Human–Computer Interaction
03
PAPER · Generative artificial intelligence, education, scaffolding strategy, self-directedness2025

Impact of Generative Artificial Intelligence on Learning: Scaffolding Strategies and Self-Directed Learning Perspectives

김태영

Generative artificial intelligence (GenAI) has penetrated various industries, including education. However, little research has studied the impact of such learning experiences in education. This study measured how the learning experiences of users varied depending on the type of scaffolding educational strategy provided by GenAI, and it identified whether learners’ self-directedness moderated the effect of GenAI on learning satisfaction. It was determined that learning flow and self-efficacy increased learning satisfaction when GenAI was used with cognitive scaffolding methods. In addition, the effect of learning flow on learning satisfaction was greater for learners with strong self-directedness, while the effect of self-efficacy on learning satisfaction was more relevant for learners with weak self-directedness.

International Journal of Human–Computer Interaction
04
PAPER · 인간-AI 상호작용, 게이밍, 토픽 모델링, 감성 분석, 사용자 경험, 기대 일치 이론, Human-AI Interaction, Gaming, Topic Modeling, Sentiment Analysis, User Experience, Expectation Confirmation Theory2025

AI as Competitor or Assistant in Gaming : Analyzing Viewer Engagement and Sentiment Patterns

김태영

인공지능은 게임 산업에서 경쟁자 또는 조력자의 역할을 수행하며, 시청자 경험에 각각 다른 영향을 미친다. 본 연 구는 인간-AI 게임을 다룬 309개의 유튜브 영상을 분석하여, AI 역할이 시청자 감정과 참여 수준에 미치는 영향을 탐구하였 다. 감성 분석과 토픽 모델링 결과, 경쟁자 AI는 고도화된 전략과 전술을 통해 시청자들에게 강한 긍정적 감정을 유발했으며, 콘텐츠 제작자(유튜버)들은 중립적이고 신중한 반응을 보였다. 반면, 조력자 AI는 콘텐츠 제작자들로부터 긍정적인 평가를 받 았지만, 시청자들은 기대에 미치지 못한 성능에 상대적으로 비판적인 반응을 보였다. 본 연구는 AI 역할이 시청자 감정과 참 여 양상에 미치는 영향을 실증적으로 분석함으로써, 게임 내 AI 설계와 사용자 경험 최적화를 위한 이론적･실질적 시사점을 제시한다. 향후 연구에서는 게임 이외 다양한 분야의 AI 역할을 탐구함으로써 인간-AI 상호작용에 대한 보다 폭넓은 이해를 도모할 수 있을 것이다.

대한산업경영학회 / 산업융합연구
05
PAPER · 메타버스 창업 생태계, 평가 기준, AHP, 비즈니스 연속성, 정책 지원2026

메타버스 창업 생태계의 정책 지원 대상 선정을 위한 평가 모형 구축: AHP 기반 메타버스 특화 요인 분석

손호성

2025년 메타버스 산업은 글로벌 규모 약 560조 원으로 급성장하고 있으나, 국내 메타버스 기반 창업기업의 정책 지원 평가를 위한 주요 요인 도출이 부재한 상황이다. 이에 본 연구에서는 AHP(Analytic Hierarchy Process) 기법을 활용하여 국내 메타버스 창업 정책 지원 대상 선정을 위한 평가 요인의 상대적 중요도 분석을 기반으로 핵심 요소 도출을 진행하고자 한다. 기존 사례 문헌 연구를 통해 제1계층 4개 요인과 제2계층 14개 세부 요인으로 구성하였다. 이를 기반으로 기술보증기금 소속 창업 평가 전문가 36명을 대상으로 설문 조사를 실시하였으며, 일관성 지수 (Consistency Ratio, CR) 0.1 미만 기준을 충족한 11명의 유효 응답을 분석에 활용하였다. 최종 요인 분석 결과, 제1계층에서는 사업 안정성 (0.28235)이 가장 높은 가중치로 나타났으며, 제2계층 종합 가중치 분석에서는 메타버스 사업 이해도 (0.10923), 가상자산 수익 모델 (0.10756), 메타버스 투자 매력도 (0.10555) 순으로 중요도가 도출되었다. 이는 앞으로 메타버스 창업기업 정책 지원 평가를 위한 실무적 기준 요소로 활용될 수 있을 것이다.

한국융합학회 / 미래기술융합논문지
06
PAPER · 2026

In-Silico Sensemaking of Algorithm Aversion in an Emergency Department Digital Twin: Effects of Explainable AI Explanation Modality and Clinician Expertise on Triage Reassessment

손호성

Background: Emergency department (ED) triage relies on visible indicators and can miss occult injuries such as delayed epidural hemorrhage. Patient digital twins that integrate brain–computer interface (BCI) signals and deep biosignals can detect such occult deterioration in real time and recommend a severity upgrade; however, these recommendations frequently collide with clinicians’ heuristics and elicit algorithm aversion. How a clinician makes sense of, and ultimately accepts or rejects, a counter-heuristic artificial intelligence (AI) alert is a central cognitive-ergonomics and decision-support problem that remains poorly understood. Objective: This study examined how the explanation modality of an explainable AI (XAI) decision aid (causal narrative vs counterfactual) and clinician expertise (novice vs expert) jointly shape epistemic trust, the sensemaking trajectory, and triage reassessment in an ED digital twin, and it derives an expertise-adaptive human–machine interface (HMI) design. Methods: We conducted an in-silico persona experiment in which a large language model (LLM) was prompted with emergency-medicine novice and expert personas. Using a 2 (expertise) × 2 (explanation modality) full-factorial design with 30 replications per cell (N=120 Chain-of-Thought reasoning traces), we combined quantitative indicators (acceptance rate, number of reasoning steps to conflict resolution, and frequency of epistemic-trust markers) with constructivist grounded-theory coding. Personas were specified by expertise and experience only—cognitive bias was treated as an observed outcome rather than an injected instruction—the stimulus was held constant, and the base model, temperature (0.7), and per-run seed rotation were fixed so that the 120 traces represent a response distribution to an identical stimulus. Results: A robust expertise × explanation interaction emerged. For experts, counterfactual explanations raised acceptance markedly relative to causal narratives (50.0% vs 76.7%) and shortened conflict resolution (6.5 vs 4.4 reasoning steps), whereas for novices the two modalities differed little (86.7% vs 80.0%). Three core categories organized the sensemaking process: clinical heuristic dominance, sensemaking reconfiguration, and cognitive conflict-resolution strategies. Counterfactual explanations appeared to disrupt experts’ confirmation loop (“what would change if I ignored this signal?”), whereas novices showed high but potentially uncritical acceptance. Conclusions: In high-stakes ED decision support, algorithm aversion is better understood as a problem of mind perception and the redistribution of agency than as a fixed trust trait. XAI explanation modality operates as a cognitive-ergonomic lever whose effect is moderated by expertise, motivating expertise-adaptive HMI and autonomy-preserving governance. Because the evidence is generated in silico, the quantitative estimates should be read as a structured set of hypotheses for subsequent human validation rather than as confirmatory human findings. Clinical Trial: Not applicable. This study is an in-silico simulation using large language models and does not involve human participants or actual patient interventions.

JMIR Human Factors