Maria Poli, Fotini Maniou, I. Maniou et al. (5 authors) • Semantic Scholar
Finding: Industrial product design mediates between AI capability and human adoption through five functions: needs framing, legibility of AI behavior, trust/control calibration, ethical lifecycle embedding, and AI-as-design-tool—arguing design decides *where AI should not be used* as much as where it should.
Why it matters: The trust calibration, legibility, and AI-as-tool frameworks are directly useful for HCI teaching and for understanding how AI reshapes design practice and human-AI collaboration.
Method: Conceptual integrative narrative review synthesizing design research, HCI, computational design, and AI ethics literature; closes with five propositions for empirical testing.
Jargon: *Legibility* — how well a system's behavior is understandable to users through its physical/interactive form; *AI ethics in product lifecycles* — considering environmental and social impact of AI-embedded products from manufacture to disposal.
Finding: Overreliance on AI stems from cognitive laziness and algorithm worship, which erode critical thinking across age groups during AI-assisted information seeking and decision-making.
Why it matters: Directly relevant to behavioral science (automation bias, cognitive offloading) and AI's impact on human cognition and decision-making, though the treatment is surface-level.
Method: Literature review built around Salvi et al.'s GPT-4 persuasion experiment as an anchor case.
Jargon: *Automation bias* — tendency to over-rely on automated system outputs; *algorithm worship* — uncritical deference to AI judgments as inherently superior to human judgment.
Note: This is a narrative literature review of modest rigor — it synthesizes existing theory rather than generating new evidence, and the framing is prescriptive rather than empirical. The relevance guidance suggests the user favors more rigorous empirical or theoretical work; this paper's quality ceiling limits its usefulness despite being topically adjacent to AI's impact on cognition and behavior change.
Vivien Gergye, Erzsébet Printz-Markó, Boglárka Eisinger-Balassa • Semantic Scholar
Finding: Households underperform on sustainable housing policies due to behavioral mechanisms (present bias, mental accounting, framing effects) that prioritize short-term affordability over long-term environmental/financial benefits.
Why it matters: Touches on behavioral economics and nudge/choice architecture concepts, but applied narrowly to housing finance policy rather than HCI, learning, creativity, or team dynamics.
Gessé Evangelista, Tatiana Alencar, M. Sami et al. (5 authors) • Semantic Scholar
Finding: A web-based LLM-powered tool that auto-generates structured UX measurement plans (metrics, tools, timing, rationales) from feature descriptions, evaluated qualitatively with six practitioners.
Why it matters: Relevant to HCI/UX design practice and the challenge of connecting design decisions to measurable outcomes, with secondary relevance to AI augmenting design workflows.
Method: Exploratory qualitative evaluation with six practitioners (small N, no formal validation).
Jargon: *Leading indicators* = metrics that predict future outcomes (e.g., engagement rate); *lagging indicators* = metrics that reflect past outcomes (e.g., revenue); *operationalization* = translating abstract concepts into measurable variables.
⚠️ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Lan Zhang, Zhuo-Yan Dai, Kuo-Hsun Wen • Semantic Scholar
Finding: AR-assisted decision-making on AI-generated content (AIGC) platforms can reduce decision fatigue and improve user satisfaction when artists select among multiple AI outputs, evaluated through Cognitive Load Theory and the Technology Acceptance Model.
Why it matters: Touches on HCI, AI tools for creative work, and decision-making under cognitive load — all relevant areas — but the execution is shallow and applied narrowly to AR interface design for art selection.
Jargon: *AIGC* = AI-Generated Content platforms; *TAM* = Technology Acceptance Model, a framework predicting user adoption based on perceived usefulness and ease of use; *Web AR* = augmented reality delivered through a browser without dedicated apps.
Note: This is a theoretical/conceptual paper with no empirical study reported — the "findings" are design propositions, not measured outcomes. The writing signals low methodological rigor, and the AR-for-art-selection framing is niche enough that it likely won't be useful for teaching core HCI, creativity, or behavioral science concepts.
E. Alomar, Christopher Engelbart, Stephen Pachucki et al. (5 authors) • Semantic Scholar
Finding: A gamified platform (Refactoria) for teaching code refactoring improved student engagement and perceived usefulness in learning code smell identification, based on a 30-student classroom experiment.
Why it matters: Relevant to gamification/serious games and learning sciences, but narrowly focused on software engineering education rather than the professor's broader HCI, creativity, or behavioral science interests.
Method: Small classroom experiment (n=30) measuring perception, effectiveness, and usefulness of the gamified tool via self-report.
Jargon: *Code smells* = symptoms of poor code design; *refactoring* = restructuring code without changing its external behavior; *antipatterns* = common but counterproductive coding patterns.
⚠️ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract