Howard Kim

Prof. Dr. Howard Kim

College of AI Convergence, Seoul Cyber University, Seoul, Republic of Korea
orcid

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

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

Representative Publications

View All Publications
11 June 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
Howard KimHoward Kim
Howard Kim
SLSung-Tae Lee
Sung-Tae Lee
JLJongwon Lee
Jongwon Lee
Grading open-ended student work consistently remains a persistent challenge in higher education, and the recent rise of large language models (LLMs) has renewed interest in rubric-guided automated scoring. However, a key gap remains: most studies report correlation rather than agreement, rarely benchmark models against a local human–human baseline, and seldom test whether simple post hoc calibration improves operational fit. This study addresses that gap by examining whether a rubric-guided LLM can approximate local human grading practice for text-based responses in three university courses, using agreement-oriented rather than correlation-only evidence. A total of 930 student responses from Prompt Engineering, Photoshop Design, and AI Video Production were scored by two human raters and by ChatGPT using the same five-criterion analytic rubric (Accuracy, Logical Flow, Specificity, Quality, and Originality; 0.0–3.0 each; Total 0–15). Human consensus (HC) was defined as the mean of the two human scores and was treated as a pragmatic reference rather than a ground truth. Pairwise agreement among H1, H2, AI, and HC was evaluated using ICC(3,1), Pearson correlations, mean absolute error (MAE), Bland–Altman bias and limits of agreement (LoA); a course-specific held-out calibration analysis was additionally conducted. For the Total score, human–human agreement was strong (ICC = 0.819 [0.797, 0.839]). AI–H1 and AI–H2 Total-score agreement were ICC = 0.700 [0.666, 0.732] and 0.767 [0.739, 0.792], respectively, while AI–HC agreement was ICC = 0.763 [0.735, 0.789], with MAE = 1.603 and LoA = [−4.246, 4.045]. At the trait level, AI–HC ICCs exceeded H1–H2 ICCs for all five rubric dimensions, although Quality remained weakly defined in the human baseline. On a 70/30 held-out test split, a course-specific linear calibration modestly improved Total-score ICC from 0.774 to 0.782 and reduced MAE from 1.624 to 1.215, narrowing the LoA from [−4.290, 4.188] to [−3.157, 3.329]. However, threshold-adjacent agreement remained imperfect after calibration. The principal contribution is a conservative, multi-metric agreement benchmark of rubric-guided LLM scoring against a local human baseline, together with a held-out calibration test that informs deployment. The findings concern written responses only and support a conservative conclusion: rubric-guided LLM scoring can assist human grading under fixed local rubrics, but the current evidence supports calibrated human–AI co-grading rather than unsupervised replacement.

30 June 2015 in International Journal of Control and Automation
Verified
Journal Article
Closed access
Howard KimHoward Kim
Howard Kim
HKHyun-Suh Kim
Hyun-Suh Kim
CJChul-Ho Jung
Chul-Ho Jung
IYInhwan Yoon
Inhwan Yoon

31 October 2015 in International Journal of Multimedia and Ubiquitous Engineering
Verified
Journal Article
Open access
HKHyun-Suh Kim
Hyun-Suh Kim
Howard KimHoward Kim
Howard Kim
IYInhwan Yoon
Inhwan Yoon
CJChul-Ho Jung
Chul-Ho Jung

2 July 2025
Verified
Conference Paper
Open access
PDF file
Howard KimHoward Kim
Howard Kim
JYJaehyun Yoon
Jaehyun Yoon
SLSeung Man Lee
Seung Man Lee
SCSung Min Cho
Sung Min Cho
YKYeon Ju Kim
Yeon Ju Kim
JLJongwon Lee
Jongwon Lee
Adult learners in remote university programs often struggle with engagement and satisfaction due to limited real-time interaction and support. This presentation introduces an AI-powered learning tutor that leverages ChatGPT-4 in a context-restricted manner, operating strictly within the instructor’s lecture materials and spoken content. By grounding the AI’s knowledge solely in the course textbooks, slides, transcripts, and other instructor-provided materials, the system delivers domain-specific support that is aligned with the curriculum and instructor’s intent. This controlled knowledge scope is a key innovation: it mitigates the usual pitfalls of generative AI (such as misinformation or irrelevant “hallucinations”) by ensuring answers remain faithful to the course content. The AI tutor functions as a virtual teaching assistant, available 24/7 to answer questions and engage in dialogue about the lecture topics. Adult learners can thus seek help at any time without fear of judgment, asking questions they might hesitate to pose to a human instructor. This on-demand, non-judgmental support has been observed to boost learner confidence and promote deeper exploration of course material. Early classroom deployments of the tutor have yielded increased student engagement and satisfaction: the AI’s ability to hold extended conversations within the bounds of the courseware has enhanced learners’ cognitive engagement and the sense of an instructor presence in remote classes. Learners reported feeling more supported and motivated, and instructors noted students coming to class better prepared. We also discuss the pedagogical implications of constraining an AI tutor’s knowledge scope. This approach empowers educators to maintain content control and uphold academic standards while still harnessing AI’s scalability and responsiveness. Instructors can focus on higher-order mentoring as routine FAQs are handled by the AI, illustrating a new model of human–AI collaboration in education. Our findings highlight that a lecture-grounded ChatGPT-4 tutor can significantly enhance engagement and satisfaction among adult learners in remote settings, offering a promising pathway to more interactive, inclusive, and effective online learning experiences.

1 October 2026 in SocArXiv
Verified
Preprint
Open access
Howard KimHoward Kim
Howard Kim
KCKeun Tae Cho
Keun Tae Cho
Large language model (LLM) personas are proposed as survey respondents, yet validation outside English-speaking contexts is scarce. We evaluate how well a Korean synthetic persona panel used to condition Gemini 3.5 Flash and EXAONE reproduces digital and artificial intelligence (AI) service-use distributions of the Korea Media Panel Survey. About 8,000 personas per model answered eight service-use items and eight attitudinal constructs; responses were compared with weighted survey estimates. The overall mean absolute error (MAE) was 14-19 percentage points (pp), with binary item-mean correlations of 0.70-0.91 across waves. Segment error across five axes was 14-18 pp, with between-group signed-error ranges of 49.6/34.7 pp (Gemini/EXAONE; 39.5/31.2 without the non-comparable teen cells). Errors were model-specific: an age stereotype (Gemini) versus an acquiescence-consistent level bias (EXAONE). Generative-AI overestimation was consistent with temporal misalignment; short-form underestimation was framing-sensitive and persisted under randomized order (both shown for Gemini). Post-hoc holdout calibration on 30% of the real data, with the correction form selected inside the calibration set, cut cell MAE from 18.3/15.1 to 4.9/4.4 pp, yet direct estimation from that subsample was more accurate than the calibrated panel (3.6 pp), a synthetic-informed shrinkage estimator beat its real-only counterpart by at most 0.7 pp, and the correction did not transfer competitively across waves. The calibrated panel kept an advantage only below roughly 250-860 real responses (at most 2.3 pp over a real-only shrinkage estimator) or, for one model, on unobserved segments. In this setting, synthetic panels are not survey substitutes; their value is diagnostic.

31 March 2022 in Journal of Digital Art Engineering and Multimedia
Verified
Journal Article
Open access
Howard KimHoward Kim
Howard Kim
DKDong Ho Kim
Dong Ho Kim
정확한 컬러재현 품질 확보를 위한 디스플레이 특성화 방법