TPACK-Guided SWOT Framework

• TPACK-guided SWOT framework is a hybrid model that combines educational content, pedagogy, and technology analysis to assess internal and external factors.
• It systematically distinguishes strengths and weaknesses from opportunities and threats by interpreting them through TPACK domains instead of generic criteria.
• Future developments focus on integrating AI literacy, prompt-crafting, and ethical reasoning to ensure technology use is pedagogically meaningful and context-sensitive.

Searching arXiv for the cited papers to ground the article in the referenced literature. A TPACK-guided SWOT framework is a strategic analytic approach that combines SWOT analysis with Technological Pedagogical Content Knowledge so that internal strengths and weaknesses, and external opportunities and threats, are interpreted through technological, pedagogical, and content-related knowledge demands rather than through generic adoption criteria alone. In the literature, this hybrid has been used to study MOOC-based collaborative learning activities, K-12 Video Generative AI, AI chatbots in physics teacher preparation, computer science teaching, mathematics teacher education, multi-agent instructional workflow design, and teacher-led evaluation of AI tools, with a recurring emphasis on context-specific judgment, content-sensitive pedagogy, and the conditions under which technology use becomes educationally meaningful (Bakharia, 2017, Lee et al., 11 Mar 2025, Mohammadipour, 20 Jul 2025, Doukakis et al., 2021, Xie et al., 15 Mar 2025, Sun et al., 13 May 2026, Demszky et al., 23 Mar 2026).

1. Conceptual architecture

Within this framework, SWOT contributes a strategic distinction between internal and external factors, while TPACK specifies the kinds of knowledge that make those factors educationally significant. The physics teacher-preparation study makes this structure explicit: strengths are treated as internal affordances of AI use within course activities; weaknesses as internal limitations of AI and user interaction; opportunities as external conditions or future-facing possibilities; and threats as external risks and systemic constraints. The distinctive contribution of the hybrid is that each quadrant is read through TPACK domains rather than as a generic administrative audit (Mohammadipour, 20 Jul 2025).

The resulting logic is not simply technological evaluation. It is an evaluation of whether a technology can be aligned with instructional planning, scaffolding, critical evaluation, reflection, information seeking, conceptual understanding, assessment, and subject-specific representation. In this sense, the framework is best understood as a way of deciding how technologies should be integrated, not merely whether they are useful (Mohammadipour, 20 Jul 2025).

SWOT quadrant	Strategic meaning	Typical TPACK emphasis
Strengths	Internal affordances	TK, PK, TPK, TCK, PCK, TPACK
Weaknesses	Internal limitations	CK, TCK, PCK, sometimes TPK
Opportunities	External enabling conditions	TPK, TPACK
Threats	External risks and constraints	TPACK with ethical and cybersecurity overlays

A related conceptual development appears in work on teacher evaluation of AI tools, where TPACK is coupled with deliberative agency. That literature defines “deliberative sensemaking” as “evaluative reasoning in which teachers draw on their values and professional knowledge to weigh trade-offs and arrive at context-sensitive judgments about AI tool use,” and argues that TPACK and agency operate in a mutually reinforcing cycle: knowledge-building enables more grounded evaluative judgment, while the act of constructing criteria deepens teachers’ understanding of tools (Demszky et al., 23 Mar 2026). This places the TPACK-guided SWOT framework within a broader theory of professional judgment rather than simple adoption.

2. Knowledge domains and analytical units

The TPACK component of the framework follows the standard seven-domain structure. TK is knowledge of technological tools and how to use them; PK is knowledge of teaching and learning processes, classroom management, methodology, assessment, course planning, and educational scenarios; CK is knowledge of the subject matter itself. Their dyadic intersections are PCK, TCK, and TPK, and the integrated domain is TPACK, defined as the capacity to represent concepts through technology, use pedagogical techniques that leverage technology constructively, understand what makes concepts difficult or easy, recognize learners’ prior knowledge, and use technology to build or strengthen understanding (Doukakis et al., 2021).

Recent AI-focused work extends these domains. In mathematics teacher education, the framework is operationalized through AI-TK, AI-TCK, AI-TPK, and AI-TPACK, alongside self-efficacy and teaching beliefs. The final structural model reported there indicates that AI-TK contributes indirectly, via AI-TPK and AI-TCK, rather than directly, to integrated AI-TPACK:

$$\begin{aligned} AI\text{-}TPACK &= 0.19 \cdot AI\text{-}TPK + 0.79 \cdot AI\text{-}TCK + \varepsilon_4 \end{aligned}$$

with the direct path from AI-TK to AI-TPACK removed as insignificant (Xie et al., 15 Mar 2025). This places content-specific AI integration at the center of the framework rather than generic tool familiarity.

The physics literature extends the model further by adding AI literacy, prompt-crafting competence, epistemic verification protocols, digital fluency, ethical reasoning, Zone of Proximal Development, and cybersecurity dimensions. In that formulation, classical TPACK is necessary but no longer sufficient for AI-rich education. Teachers must understand that LLM outputs are probabilistic, that they can hallucinate or oversimplify, that prompt quality shapes response quality, and that outputs must be triangulated with textbooks, authoritative sources, peer discussion, and critical questioning (Mohammadipour, 20 Jul 2025).

Behavioral work on multi-agent instructional workflow design pushes the framework from self-report to observable enactment. There, AI-TPACK is treated as an emergent, dynamic capability manifested through design processes and shaped by systems thinking, pedagogical beliefs, self-efficacy, and scaffolded workflow-design behavior. The coded analytic dimensions are AI-TK, AI-TPK, AI-TCK, and AI-TPACK integration, but the substantive claim is that effective integration emerges from interaction among architecture, prompts, tools, branching logic, and pedagogical aims rather than from the possession of isolated domains alone (Sun et al., 13 May 2026).

3. Internal strengths and weaknesses

A core use of the framework is to diagnose internal capability profiles. In upper secondary computer science teaching, a national sample of in-service teachers produced a highly differentiated TPACK profile: CK $M = 4.38$, TPK $M = 4.18$, TK $M = 4.16$, PK $M = 4.12$, TPACK $M = 4.03$, TCK $M = 3.68$, and PCK $M = 3.51$. The strongest correlation was $r(\text{TK},\text{PK}) = 0.746$, the second strongest was $r(\text{TPK},\text{TPACK}) = 0.715$, and the weakest was $M = 4.38$0. In SWOT terms, this profile indicates strong content, technology, and general pedagogy, but weaker content-linked intersections, especially pedagogical transformation of domain knowledge and technology-content alignment (Doukakis et al., 2021).

The mathematics teacher-education study reports a different but structurally similar pattern. AI-TPACK among mathematics teacher education students is at a “basic, preliminary” or “nascent” stage; Teaching Beliefs has the highest mean ($M = 4.38$1), AI-TCK is the highest among the AI-TPACK components ($M = 4.38$2), and AI-TK is the lowest ($M = 4.38$3). Self-efficacy positively predicts AI-TK ($M = 4.38$4), AI-TCK ($M = 4.38$5), and AI-TPK ($M = 4.38$6), whereas stronger teaching beliefs are associated with lower AI-TPK and AI-TCK in the final model (Xie et al., 15 Mar 2025). This gives the framework a psychologically differentiated internal structure: self-efficacy functions as an asset, while rigid or excessive teaching beliefs may function as a barrier.

In K-12 Video GenAI, teachers identified internal strengths that are largely TCK-, TPK-, and TPACK-related: dynamic visualization of hard-to-teach content, support for engagement, promotion of creativity and divergent thinking, personalized and differentiated instructional materials, authentic task design and assessment, and improved efficiency in lesson preparation. Internal weaknesses were equally specific: output instability and limited controllability, weak content accuracy and risk of misconception, incomplete support for novice prompting, language limitations, fragmented educational workflow, and the need for heavy teacher mediation (Lee et al., 11 Mar 2025).

The physics chatbot study sharpens the same pattern in a symbolically dense STEM domain. Reported internal strengths include pedagogical translation of complex concepts, enhanced information-seeking and exploratory prompting, support for symbolic reasoning, metacognitive reflection and iterative improvement, productivity and planning support, and assistance with simulation and code-related planning such as Python debugging and Arduino-based experimental design. Internal weaknesses include domain-specific inaccuracies, symbolic limitations, incorrect LaTeX formatting, pedagogical misalignment, need for prior expertise to detect errors, and risks of overreliance and de-skilling. The malformed expression

$M = 4.38$7

is used in that study as an example of faulty symbolic output (Mohammadipour, 20 Jul 2025).

4. External opportunities and threats

The opportunity side of the framework identifies enabling conditions beyond the immediate user-tool interaction. In the physics literature these include inclusive and multilingual education, expansion of the Zone of Proximal Development through just-in-time guidance and alternative explanations, development of AI literacy and digital competence, curriculum-level integration and teacher education reform, and professional reflection on trustworthiness and professional identity. At the institutional level, the same literature points to secure infrastructure, formal guidance and policies, equitable access, professional development, and sustainable or open-source AI options as strategic enablers (Mohammadipour, 20 Jul 2025).

K-12 Video GenAI research identifies closely related opportunities: multimodal curriculum innovation; authentic project-based learning and student production; integration of AI literacy, media literacy, and ethics education; professional development targeted to Video-GenAI-specific TPACK; school and system-level support through clearer guidelines, funding, infrastructure, and technical support; and better educational tool design through teacher dashboards, multilingual prompting, safer filters, easier sharing, and stronger consistency control (Lee et al., 11 Mar 2025). In that study, TAM is used alongside TPACK, and the relation to SWOT is explicit: strengths are often reinforced by perceived usefulness, weaknesses often arise from low or unstable perceived ease of use, opportunities depend on improving usefulness and ease of use through training and support, and threats can suppress adoption even when usefulness is high (Lee et al., 11 Mar 2025).

The threat side of the framework is unusually prominent in AI-related work. In physics teacher preparation, threats include prompt injection and adversarial manipulation, institutional access gaps and inequality, privacy and platform vulnerabilities, instructor preparedness gaps, sustainability and licensing issues, and systemic overuse that erodes initiative and symbolic fluency (Mohammadipour, 20 Jul 2025). In Video GenAI, external threats include harmful or deceptive content, bias and amplified misconception, privacy and copyright uncertainty, lack of guidelines, equity gaps, age restrictions, cost, infrastructure disparities, and administrative burdens such as consent procedures and alternative arrangements for non-consenting students (Lee et al., 11 Mar 2025).

In mathematics teacher education, the threat environment is more structurally pedagogical. Reported risks include entrenched traditional teaching beliefs, exam-oriented and time-constrained mathematics teaching culture, informal and unsystematic Internet-based AI learning, weak AI modeling during internships, and institutional inertia signaled by the absence of meaningful grade-level growth in AI-TK, AI-TCK, AI-TPK, and AI-TPACK (Xie et al., 15 Mar 2025). In upper secondary computer science, time pressure, large class sizes, inadequate laboratories, lack of educational materials and scenarios, and a paper-based national examination similarly constrain enactment even where teacher confidence in TK, PK, CK, and TPK is high (Doukakis et al., 2021).

Work on deliberative tool evaluation adds another category of threat: exclusion of teachers from adoption decisions. That literature highlights bias, stereotype reinforcement, unequal subgroup treatment, cognitive offloading, expertise erosion, opaque model behavior, and top-down procurement based on compliance, cost, or scalability rather than pedagogical fit. It therefore treats agency loss itself as a systemic threat condition (Demszky et al., 23 Mar 2026).

5. Pedagogical scripting, workflow, and evidence practices

The framework is not limited to evaluative matrices; it also governs activity design. In MOOC research, SWOT is explicitly treated as a “multi-perspective elaboration” activity rather than a free-form discussion thread. An instructor can create a SWOT template as a multi-perspective fieldset containing Strengths, Weaknesses, Opportunities, and Threats; learners may choose a perspective, contribute to all perspectives, or be randomly assigned a perspective; and the resulting submissions are stored in a searchable knowledge base that may persist across course re-runs. The tool design further includes optional curation stages, moderator highlighting, natural language processing, Latent Dirichlet Allocation for topic modeling, clustering of similar responses, and scalable feedback at the level of clusters rather than individual submissions (Bakharia, 2017).

That workflow changes the meaning of SWOT in educational settings. It becomes a scripted sequence of individual contribution, knowledge-base construction, peer exploration, curation, moderation, and feedback. The knowledge base is not a transient forum but a persistent communal resource. In TPACK terms, CK is encoded through explicit perspective categories; PK through collaborative knowledge construction, reflection, and curation; and TK through searchable storage, persistence, moderation, and NLP-supported summarization (Bakharia, 2017).

Teacher-evaluation research adds a complementary process model. Rather than beginning with fixed quadrants, it documents five mechanisms that support high-quality evaluative reasoning: time and space for deliberation, artifact-centered sensemaking, collaborative reflection through diverse viewpoints, knowledge-building, and psychological safety. In that setting, teachers generated 211 criteria organized into four higher-order themes—Practical, Equitable, Flexible, and Rigorous—and nearly equal attention was given to student-facing outputs and teacher-facing process burdens. This suggests that a TPACK-guided SWOT framework functions best when it is enacted as deliberative sensemaking rather than as a rapid checklist (Demszky et al., 23 Mar 2026).

Behavioral work on multi-agent design demonstrates how such frameworks can be linked to observable practice. Using 8,718 logged actions from 61 teachers, that study identified three archetypes: Systematic Optimizers, Prolific Creators, and Passive Observers. Systematic Optimizers showed iterative refinement, configuration-intensive work, and stronger AI-TPACK coherence; Prolific Creators rapidly produced practical tools but often relied on scaffolds and feature stacking; Passive Observers were browsing-dominant and showed polarized expert-novice outputs. The study therefore recommends differentiated scaffolding: structured templates and real-time assistance for weaker Passive Observers, pedagogical frameworks and exemplar repositories for Prolific Creators, and advanced documentation and open architectural affordances for Systematic Optimizers (Sun et al., 13 May 2026).

6. Limitations, misconceptions, and future development

A recurrent misconception is that the framework itself demonstrates effectiveness. The literature does not support that conclusion. The MOOC study is explicitly a work-in-progress design paper with mockups and design principles rather than a validated intervention, and it does not present learning outcome data, usability testing results, comparative evidence against forums, or live-course implementation results (Bakharia, 2017). The physics chatbot study is bounded to a single course context and a qualitative dataset based on reflections, observations, and discussion transcripts (Mohammadipour, 20 Jul 2025). The mathematics teacher-education study relies on survey data and self-report, with the authors noting the need for interviews, videos, and multimodal data (Xie et al., 15 Mar 2025). The Video GenAI study examined the perceptions of 10 purposefully selected “leading teachers,” not student outcomes or representative teacher populations (Lee et al., 11 Mar 2025). The multi-agent workflow study occurred in a two-day workshop on a specific no-code platform and does not provide direct classroom implementation evidence (Sun et al., 13 May 2026).

A second misconception is that a TPACK-guided SWOT framework is equivalent to a generic checklist of benefits and risks. The literature instead treats it as a theory-anchored diagnosis of how technologies interact with disciplinary representation, pedagogical orchestration, teacher judgment, and institutional conditions. In AI-rich settings, the framework is repeatedly extended with AI literacy, prompt-crafting, verification routines, ethical reasoning, cybersecurity awareness, systems thinking, and self-efficacy because classical TPACK alone does not capture probabilistic outputs, hallucination, prompt sensitivity, or prompt injection risks (Mohammadipour, 20 Jul 2025, Sun et al., 13 May 2026).

A third misconception is that technology knowledge alone is sufficient. Across the surveyed work, the more consequential issues arise at the intersections: PCK and TCK in computer science, AI-TCK and AI-TPK in mathematics teacher education, TCK and PCK in Video GenAI, and symbolic accuracy plus pedagogical alignment in physics. This suggests that the strategic value of the framework lies precisely in preventing technology adoption from being reduced to tool exposure or interface familiarity (Doukakis et al., 2021, Xie et al., 15 Mar 2025, Lee et al., 11 Mar 2025, Mohammadipour, 20 Jul 2025).

Future development in this literature is therefore oriented toward deeper integration rather than broader enthusiasm. Recurring priorities include structured and bounded activities, verification protocols, prompt-crafting instruction, AI ethics, digital fluency training, secure infrastructure, multilingual and multimodal support, adaptive analytics, stronger teacher dashboards, curriculum redesign, differentiated professional development, and institutional guidance that treats teachers as agentic evaluators rather than implementation targets (Mohammadipour, 20 Jul 2025, Lee et al., 11 Mar 2025, Xie et al., 15 Mar 2025, Demszky et al., 23 Mar 2026). The cumulative implication is that a TPACK-guided SWOT framework is most robust when it serves as a strategic, epistemically aware, and context-sensitive model for diagnosing capacities, constraints, and safeguards in educational technology integration.