Research Problems
This chapter synthesizes research problems, gaps, and limitations identified through the Systematic Literature Review (SLR) of 55 papers on AI and critical thinking in business education (ScholarRAG, 2026-04-09).
Note: This analysis was generated as part of the ScholarRAG SLR pipeline. All identified gaps and problem statements have been verified by the research team.
Section 1: Research Problem Landscape
Problem Theme 1: Inconsistent Evidence on AI’s Effect on Critical Thinking
🔴 Critical
Empirical studies on the effect of AI tool use on students’ critical thinking (CT) in business education produce conflicting results, with no theoretical or methodological consensus. Essien et al. (2024) found that ChatGPT use showed meaningful improvement only at lower Bloom levels (remembering, understanding, applying), with limited effects on higher-order thinking. Gerlich (2025) reported null results (p > 0.05), warning that AI integration does not automatically guarantee CT improvement. Together, these studies signal that AI tool adoption is necessary but not sufficient — pedagogical design and contextual conditions are the key determinants of CT effects (Valcea et al., 2024).
Problem Theme 2: Intended–Enacted Curriculum Gap and Institutional Policy Lag
🔴 Critical
Even when business schools explicitly claim AI–CT integration in official documents (Intended Curriculum), a structural gap exists between those claims and instructors’ actual teaching practice (Enacted Curriculum). Institutional policy consistently lags behind practice. Stewart et al. (2026) confirmed this qualitatively through inductive coding of 18 education leadership students’ reflective records. Nowinski et al. (2025) demonstrated through PLS-SEM (N=133) that instructors’ AI integration intentions are shaped by formal AI training and self-efficacy pathways — not mere experience accumulation — suggesting that the quality of experience, not its quantity, governs Intended–Enacted alignment.
Problem Theme 3: Heterogeneity in Student Experience (Achieved Curriculum) and AI Over-Reliance
🟡 Emerging
Student CT development in AI-integrated curricula is bifurcated by usage patterns. Albannai (2025) found (N=40 DBA students) a heterogeneous spectrum from complementary use (stimulating CT) to over-reliance (insufficient critical verification). Fischer et al. (2024) found that approximately 90% of graduate students (N=118) used AI only for surface tasks (writing, paraphrasing), with fewer than 1 in 10 achieving higher-level sensemaking. Abdelwahab et al. (2023) showed that Dutch business students widely perceived their institutions as not adequately preparing them for AI work environments, highlighting the Intended–Achieved perception gap.
Problem Theme 4: Absence of Cross-National AI–CT Integration Research
🟡 Emerging
Research on AI–CT integration in business education is concentrated in specific countries (primarily the UK, US, and India), with virtually no systematic cross-national comparative research involving the US, China, and Korea. Ode et al. (2025) conducted a UK–Nigeria comparison, showing cultural differences affect AI tool use motivation and continuance intention — but this is not specific to AACSB business schools and does not directly examine CT integration patterns. Liang et al. (2026) applied fsQCA in a Chinese logistics education context but cross-validation for the US or Korean context has not been conducted.
Problem Theme 5: Institutional Alignment Gap Between AACSB Standards and AI–CT Integration
🟢 Ongoing
Systematic empirical research examining how AI–CT integration aligns with AACSB accreditation learning outcome standards is scarce. Akhtar et al. (2024) confirmed misalignment between existing curricula and industry AI/big data competency demands in 38 AACSB supply chain programs. Desai (2024) argued conceptually that business schools can play a key role in AI talent development, but explicit empirical mapping of CT and AI integration in the AACSB accreditation context has not been conducted.
Section 2: Research Gaps
Theoretical Gaps
T1: Absence of an Integrative IEA Curriculum Theory Framework
The literature lacks an integrative theoretical framework that simultaneously encompasses Intended, Enacted, and Achieved curriculum dimensions for AI–CT integration. Individual dimensions are studied (Xu & Babaian, 2021; Nowinski et al., 2025; Abdelwahab et al., 2023), but no study theorizes the alignment and gaps between all three layers simultaneously.
T2: Absence of Comparative Education Theory Framework
Analysis of the SLR corpus shows Bloom’s Taxonomy, TAM, SDT, and TPB are repeatedly used, while comparative education theory frameworks are almost never adopted. Ode et al. (2025) combined SDT and EDT for UK–Nigeria comparison, but this is insufficient to theorize country-specific pathway differences (equifinality) in the US–China–Korea AACSB context.
T3: Underdevelopment of Complex Causal Theory Permitting Equifinality
Most literature relies on theoretical frameworks explaining AI–CT integration outcomes through single-pathway linear causality. Only Liang et al. (2026) used fsQCA to explore asymmetric causality and equifinality — limited to China. A theoretical framework permitting multiple pathways (equifinality) across countries is underdeveloped.
Methodological Gaps
M1: Absence of Longitudinal Research Design
Most studies rely on cross-sectional designs. No genuine longitudinal tracking exists to study long-term CT competency development following AI curriculum integration.
M2: Insufficient Systematic Curriculum Document Analysis
Studies conducting systematic coding and comparative analysis of actual curriculum documents are very rare. Akhtar et al. (2024) and Lyytinen et al. (2021, 2023b) conducted curriculum analysis but did not systematically code Explicit/Implicit/Nominal AI–CT integration levels for cross-national comparison.
M3: Absence of Multi-Stakeholder International Comparative Mixed Methods
Gupta and Jaiswal (2024) included both students (SEM N=525) and faculty (interviews N=35), but limited to India. No international comparative research using multi-stakeholder mixed methods (curriculum analysis + instructor interviews + student surveys) across the US, China, and Korea currently exists.
M4: Rarity of fsQCA Application in Business Education AI–CT Research
fsQCA appears in only 2 papers in the SLR corpus (Gong et al., 2025; Liang et al., 2026), both limited to China. Application of fsQCA to cross-national comparison of AI–CT Explicit integration in AACSB business schools is absent.
Contextual Gaps
C1: Complete Absence of Korean Business Education Context
Not a single paper in the SLR corpus sets a Korean AACSB-accredited business school as the research context. This completely blocks exploration of country-level pathway differences within East Asian business education.
C2: Insufficient AACSB-Specific AI–CT Integration Research
Studies explicitly targeting AACSB-accredited institutions are limited to Gupta et al. (2025), Akhtar et al. (2024), and Gupta (1994). None analyzes AI–CT integration patterns in relation to AACSB accreditation requirements.
C3: Rarity of UG–MBA Cross-Program Level Comparison
Only Lyytinen et al. (2021) simultaneously covers undergraduate and MBA levels — limited to IS curriculum without CT linkage. Studies including MBA programs do not conduct systematic comparison with undergraduate programs.
Section 3: Limitations of Existing Literature
| Structural Limitation | Description |
|---|---|
| Single-country/institution bias | Most studies from a single country (UK, India, China, US) with no cross-national comparability |
| Small samples and non-standardized measures | Many studies: N=15–40; only Gerlich (2025) uses standardized CT assessment (Watson-Glaser) |
| CT concept inconsistency | “Critical thinking” operationalized as cognitive levels (Bloom), decision-making, problem-solving, sensemaking, systems thinking, AI literacy — no theoretical consensus |
| Instructor perspective bias | Faculty-perspective studies concentrated in specific countries; no simultaneous US–China–Korea comparison |
| No IEA three-layer simultaneous study | No study examines all three curriculum layers (Intended, Enacted, Achieved) in a single international comparative design |
| Excess conceptual papers | Conceptual papers and literature reviews constitute a large share; frameworks proposed are not empirically validated |
Section 4: New Research Problem Statements
PS1: Cross-National Distribution of AI–CT Integration Clusters in AACSB Schools
Statement: How do Explicit/Implicit/Nominal AI–CT integration cluster distributions systematically differ across US, China, and Korea in AACSB-accredited business schools, and how does program level (UG vs. MBA) moderate these differences?
RQ link: RQ1, RQ1a, RQ1b | Theory: Bloom’s Taxonomy + Comparative Education Theory | Method: Mixed (curriculum content analysis + ANOVA/chi-square + NLP text analysis)
PS2: Equifinality of Conditions for Explicit AI–CT Integration by Country
Statement: What combinations of conditions lead to Explicit AI–CT integration in AACSB business schools, and do country-specific pathways (equifinality) exist across the US, China, and Korea?
RQ link: RQ2, RQ2a, RQ2b | Theory: Dynamic Capabilities Theory | Method: Mixed (fsQCA + curriculum analysis + institutional characteristics data)
PS3: Causal Structure and Cross-National Variation of the Intended–Enacted Gap
Statement: What is the nature and causal structure of the gap between AI–CT Intended Curriculum (documents) and Enacted Curriculum (actual teaching), and how does this gap differ across the US, China, and Korea?
RQ link: RQ3 | Theory: TPB | Method: Qualitative (semi-structured interviews + thematic analysis) or Mixed (interviews + fsQCA)
PS4: Cross-National and Cross-Cluster Variation in Student CT Perception and Experience
Statement: How do CT perception and experience of students enrolled in AI-integrated curricula in AACSB business schools systematically differ across the US, China, and Korea, and across Explicit/Implicit/Nominal clusters?
RQ link: RQ5 | Theory: SDT | Method: Quantitative (survey + SEM) or Mixed (survey + focus group interviews)
PS5: Cross-National Variation in Instructor-Perceived Facilitating and Inhibiting Factors
Statement: How do instructor-perceived facilitating and inhibiting factors for AI–CT integration systematically differ across the US, China, and Korea, and by what institutional and cultural conditions are these differences explained?
RQ link: RQ4 | Theory: SCOT | Method: Qualitative (semi-structured interviews + IPA or thematic analysis)
PS6: International Validation of Equifinality via fsQCA
Statement: Are there both universal conditions (operating regardless of country) and country-specific pathways (equifinality) for Explicit AI–CT integration across UG and MBA programs in AACSB schools?
RQ link: RQ2a, RQ2b | Theory: Dynamic Capabilities + Triadic Reciprocal Determinism | Method: Mixed (fsQCA + curriculum content analysis + institutional characteristics)
Gap–RQ Mapping Summary
| RQ | Primary Gaps Addressed | Supporting Gaps |
|---|---|---|
| RQ1 (Pattern comparison) | C1 + C2 + I1 + M2 | T2 + C3 |
| RQ2 (Condition combinations) | T3 + M4 + T2 | C1 + I2 |
| RQ3 (I–E gap) | T1 + M3 + C2 | C1 + I2 |
| RQ4 (Instructor factors) | T2 + M3 + C1 | C2 + I2 |
| RQ5 (Student experience) | T1 + M3 + C1 | I1 + Tp2 |
Legend: T = Theoretical gap, M = Methodological gap, C = Contextual gap, Tp = Temporal gap, I = Integrative gap