Finding: Narrative review arguing that generative AI is shifting smart education from automation toward personalization and generation, identifying five application directions and a Technology–Pedagogy–Governance governance framework.
Why it matters: Directly relevant to AI's impact on learning sciences, skill acquisition, and human-AI collaboration in educational contexts, and raises behavioral/ethical questions about human-AI collaboration models useful for HCI and AI-impact teaching.
Method: Narrative review (not systematic/meta-analytic), so findings are descriptive and framework-building rather than empirical.
Jargon: *Intelligent tutoring systems (ITS)*: adaptive software that provides personalized instruction without human intervention; *Technology layering*: the idea that successive AI capability tiers (rule-based → ML → LLM) determine what educational applications are feasible at each level.
Finding: Combining personalized goal-setting with low-cost symbolic incentives (~$1 awards) improved math test scores by 0.35 SD among disadvantaged elementary students compared to a control group receiving standard instruction.
Why it matters: Directly relevant to goal-setting research and behavioral science — specifically the intersection of incentive design, motivation, and learning outcomes — with practical implications for low-cost nudge/reward interventions in educational contexts.
Method: Quasi-experimental difference-in-differences design with 426 students; exploratory subgroup analysis for non-primary-language households.
Jargon: *Difference-in-differences*: a statistical technique comparing outcome changes over time between a treatment group and a control group to estimate causal effects; *symbolic incentives*: non-monetary rewards (e.g., certificates, trophies) that signal recognition rather than financial value.
⚠️ Single study (not replicated) | Limited statistical detail in abstract
Finding: Paper prototyping as an early-stage usability evaluation method effectively identifies usability issues (navigation, cognitive load, alert interpretation) in IoT emergency response dashboards, with 32 issues found and resolved across 15 participants.
Why it matters: Directly relevant to HCI and UX design pedagogy, particularly the value of low-fidelity prototyping and sketching in early design stages for safety-critical systems.
Method: Mixed-methods with simulated emergency scenarios, observation, structured questionnaires, and semi-structured interviews with 15 participants.
Jargon: *Paper prototyping* — hand-drawn mockups used to test interface designs before digital implementation; *cognitive load* — mental effort required to process information.
⚠️ Self-reported data only | No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Andres Fortino, Siranush Kostanyan • Semantic Scholar
Finding: AI-integrated capstone projects using Agile methodologies produced significantly higher client satisfaction and publication-worthiness than non-AI projects (45.3% vs. 29.8% "A" ratings), and the authors propose a "competitive collaboration" model reconciling individual assessment with team-based work.
Why it matters: The "competitive collaboration" design and Agile-in-education framework touch on team dynamics, AI's impact on work/learning, and skill acquisition in ways relevant to teaching HCI and behavioral science courses.
Method: Mixed-methods: two qualitative case studies plus quantitative analysis of 146 capstone projects comparing AI vs. non-AI project outcomes.
Jargon: *Agile methodologies* — iterative, sprint-based project management frameworks; *competitive collaboration* — a course design where students compete individually while working in client-facing teams.
Finding: The MAOV framework uses multimodal behavioral data to identify UX bottlenecks in smart home interfaces, showing 30-50% task time reductions and significant usability score improvements after one redesign cycle.
Why it matters: Marginally relevant to HCI/UX design methods, but the focus is narrowly on smart home engineering optimization rather than the behavioral science, creativity, team dynamics, or AI-society themes the professor emphasizes.
Method: Within-subjects experiment (N=40) with K-means++ clustering of behavioral archetypes from five synchronized data streams.
⭐ Longitudinal design ⚠️ Small sample (n=40) | No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Mehmet Fatih Özalp, Hicran Hanım Halaç, Nuray Özkaraca Özalp et al. (4 authors) • Semantic Scholar
Finding: Bibliometric analysis of 45,230 eye-tracking publications on Scopus from the 19th century to 2025, showing explosive growth 2020–2025 and increasing integration with VR/AR, EEG, ML, and deep learning.
Why it matters: Eye-tracking is peripherally relevant to HCI and UX research methods, but this paper is purely a bibliometric mapping exercise with no substantive findings about behavior, design, or interaction—limited teaching or research utility.
Method: Bibliometric analysis using Bibliometrix in RStudio on Scopus corpus (n=45,230).
Patricia Sigalingging, N. N. Aritonang • Semantic Scholar
Finding: Achievement Goal Orientation (AGO) shows no significant relationship with academic procrastination among AI-using university students (r = −0.094, p = 0.090).
Why it matters: Tangentially relevant to goal-setting and motivation research, but the null finding in a single Indonesian university sample with no meaningful AI integration into the analysis limits applicability.
Method: Spearman Rank correlation with purposive sampling (n=329), using validated scales (AGQ-R and MAPS-15).
Jargon: *Achievement Goal Orientation* = motivational framework distinguishing mastery (learning-focused) vs. performance (comparison-focused) goals; *Mastery-Approach* = orientation toward developing competence relative to one's own past performance.
Finding: Learner autonomy—students taking responsibility for their own learning through goal-setting, self-regulation, and strategy selection—positively predicts English language achievement in university students.
Why it matters: Touches on self-regulation and goal-setting constructs relevant to learning sciences and motivation research, but narrowly focused on EFL/ESL pedagogy rather than the broader behavioral science or HCI topics the user prioritizes.
Ni Made Citariani, K. Agustini, I. K. Suartama et al. (6 authors) • Semantic Scholar
Finding: An AR framework combining brain-based learning and cultural values (Tri Kaya Parisudha) improves preschoolers' prosocial, critical thinking, and communicative behavior in a quasi-experimental study.
Why it matters: Tangentially relevant to learning sciences and educational technology design, but the focus on early childhood character education and Balinese cultural values is quite narrow and unlikely to connect to HCI, creativity, teams, or behavioral science teaching at a university level.
Method: ADDIE instructional design model with expert validity review and quasi-experimental classroom evaluation.
⚠️ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Finding: A VR-based instructional system for spoken English practice showed no statistically significant improvement over traditional instruction (Cohen's d = 0.21), though learners reported high engagement and confidence.
Why it matters: Marginal relevance to HCI/immersive tech and learning sciences, but the null quantitative result with positive qualitative perception is a recurring pattern worth noting in VR-for-learning literature.
Method: Convergent parallel mixed-method design with pre/post IELTS assessments and ANCOVA across two intact classroom groups (n=60).
⚠️ Small sample (n=30) | Self-reported data only | No control group mentioned | Single study (not replicated)
Husni Charif, Nour Damaa, Sam El Nemar et al. (5 authors) • Semantic Scholar
Finding: AI adoption in Lebanese schools boosts educator productivity indirectly through engagement, with change leadership as the key driver; engagement mediates the AI-productivity relationship.
Why it matters: Tangentially relevant to AI impact on work and human collaboration, but the educational leadership/HR management framing is far from the professor's core HCI, creativity, or behavioral science interests.
Method: SEM with CFA on Lebanese educator survey data.
Jargon: *Social cognitive theory (SCT)*: Bandura's framework linking beliefs, environment, and behavior; *Dynamic capabilities theory (DCT)*: organizational ability to adapt resources in changing environments; *Confirmatory factor analysis (CFA)*: statistical method to validate measurement models.