Prompts to Practice: A Pedagogical Framework for Human-Centered AI Engagement

Introduction

The rapid integration of generative artificial intelligence (AI) into academic and professional life demands a deliberate pedagogical response from higher education, one that goes beyond policy and addresses how students learn to use these tools with intention and judgment. This manuscript focuses specifically on text-based large language models (LLMs) such as ChatGPT, Copilot, Gemini, Perplexity, and Claude. While generative AI encompasses a broader range of systems that produce images, audio, and video, LLMs serve as the primary lens here because of their direct role in academic writing, reasoning, and knowledge construction. Rather than treating these tools as threats to be prohibited, institutions have an opportunity to teach students to use them effectively, responsibly, and ethically. We describe this approach as human-centered AI engagement, which is defined in the following section.

Emerging research deepens the urgency of this challenge. Bozkurt et al. (2026) describe the emergence of the “ghost student,” a digital surrogate capable of navigating learning management systems, completing assignments, and submitting coursework without the learner’s active participation. This development reframes concerns about academic integrity not merely as a matter of individual misconduct, but as a structural vulnerability within current assessment models, one that calls for a pedagogical response rather than a disciplinary one. When generative AI substitutes for student thinking rather than supporting it, learners risk accumulating what Bozkurt et al. (2026) describe as “cognitive debt,” a condition in which the productive struggle essential to deep learning is quietly outsourced to automated systems. These developments expose the limits of reactive, compliance-focused approaches and signal the need for intentional instructional frameworks.

Recent scholarship reinforces this point. AI in education is not a transient phenomenon but a systemic transformation reshaping teaching, learning and leadership practices across educational context (Arar, Özen, et al., 2025; Mustafa et al., 2024). DeMatthews and colleagues argue that AI is no longer simply a technological innovation but a leadership imperative requiring ethical responsibility, equity-centered decision-making, and strategic guidance across educational systems (2026). While that work focuses on K-12 leadership, it strengthens the argument that higher education must also move beyond reactive responses toward human-centered frameworks for AI integration.

Within higher education, faculty attitudes toward AI vary widely. Some instructors embrace these tools for their potential to support learning and increase access, while others remain skeptical of their academic and ethical implications (Chan & Tsi, 2023; Huang et al., 2025; Joyce, 2023; Tu, 2024). This range of perspectives reflects an ongoing, unresolved debate over whether AI represents a benefit or a risk for education (Mogavi et al., 2024; Rawal, 2025; Theresiawati et al., 2025). What is increasingly clear, however, is that faculty uncertainty alone cannot substitute for purposeful pedagogical guidance.

Much of the existing literature on AI in higher education addresses policy compliance, academic integrity enforcement, or general AI literacy. These approaches share a common limitation: they position students as risks to be managed rather than learners to be developed. Less attention has been given to instructional frameworks that treat LLM engagement as a teachable, scaffolded process, one that builds ethical reasoning, rhetorical awareness, and intellectual ownership over time. Existing frameworks do address AI in education, but largely at the level of policy or competency rather than classroom enactment (e.g., Chan, 2023; Chiu & Rospigliosi, 2026; UNESCO, 2024). Chan (2023), for example, offers a comprehensive institutional framework for governing AI use in university teaching and learning, while UNESCO (2024) and the broader landscape of official frameworks synthesized by Chiu and Rospigliosi (2026) define the competencies—the “what”—that learners and educators are expected to develop. These frameworks provide essential reference points, yet they stop short of specifying how individual instructors might cultivate such competencies in day-to-day practice. The present model addresses that gap by offering a classroom-level “how”: a phase-sequenced instructional process, centered on prompt literacy, that an instructor can enact directly with students. While much of the existing scholarship has focused on the leadership, policy, and ethical implications of AI integration across educational systems (Arar, Tlili, et al., 2025), less attention has been given to how these principles are enacted within classroom pedagogy. Leadership scholarship establishes the system-level mandate for ethical, equity-centered AI integration (Arar, Özen, et al., 2025; DeMatthews et al., 2026); the framework offered here operationalizes that mandate at the instructional level, translating the values leaders articulate into concrete classroom practice. In this sense, leadership defines why human-centered AI integration matters, while the present model provides the means by which faculty can enact it. In response, this article introduces the Instructional Model for Human-Centered Generative AI Engagement, a constructionist pedagogical framework designed to guide students in engaging with generative AI as a thinking partner rather than a shortcut. A central contribution of the framework is its distinction between prompt engineering, which focuses on technical optimization of AI inputs, and prompt literacy, which emphasizes rhetorical awareness, ethical responsibility, and critical interpretation. Drawing on classroom practice and existing research, the model offers instructors practical strategies for scaffolding AI-integrated learning while keeping human thinking at the center. Central to this is the role instructors play in modeling transparency and intentional practice, reflecting the framework’s commitment to deliberate and ethical relationship with LLMs, one in which AI serves as a collaborator in learning rather than a replacement.

Higher education is a particularly important site for this work. Unlike K-12 contexts where curriculum is more tightly structured, faculty in higher education have significant autonomy over pedagogical design, making instructor-level frameworks especially actionable. College students are also at a critical stage of developing independent intellectual identity, and how they learn to engage with LLMs now will shape their professional and academic practices long after graduation.

Human-Centered AI Engagement

Central to this framework is the concept of human-centered AI engagement, which we define as an intentional, reflective approach in which students work with generative AI as a thinking partner rather than a shortcut. In this view, the goal of instruction is not to maximize what AI can produce, but to ensure that students exercise judgment, ethical responsibility, and ownership over their learning, so that human reasoning remains at the center of knowledge construction. Human-centered AI engagement therefore treats generative AI as a collaborator in thinking, one whose outputs are interrogated, revised, and integrated rather than passively accepted.

This orientation distinguishes the present approach from two common responses to generative AI in higher education. The first treats AI primarily as a threat to be prohibited or policed, an essentially reactive posture that positions students as risks to be managed. The second embraces AI for its efficiency, prioritizing speed and convenience in ways that can quietly displace the productive struggle essential to deep learning. Human-centered AI engagement instead foregrounds intentionality, human judgment, ethical awareness, and learner agency. It asks not whether students use AI, but how they engage with it: whether they do so critically, transparently, and in service of their own intellectual growth. The remainder of this article develops this orientation, first through its theoretical grounding in constructionism and then through an instructional model that translates these principles into classroom practice.

We come to this topic as educators and scholars committed to exploring the pedagogical and ethical implications of AI in higher education. With backgrounds in instructional design, college teaching, student development, and educational and community leadership, we bring practical experience and critical inquiry to this work. We approach AI integration not as a question of technology alone, but as an opportunity to reimagine teaching and learning through a constructionist lens.

A Constructionist Lens for Incorporating AI

A constructionist pedagogical approach, grounded in Papert’s (1980) work, offers a powerful lens for guiding students in ethical, responsible, and effective practices with AI. Constructionism emphasizes that learners construct knowledge most meaningfully when they create tangible, personally meaningful artifacts and engage in reflective dialogue (Bernauer & Fuller, 2017; Papert & Harel, 1991). When applied to LLMs, this means inviting students to engage as co-creators, to draft, explore ideas, and/or build digital artifacts that reflect their thinking, authorship, and values.

Rather than treating AI as a threat to academic integrity, constructionist pedagogy frames generative AI as “objects to think with,” cognitive and creative partners that spark inquiry, deepen engagement, and promote critical thinking (Adel et al., 2024; Chen et al., 2025; Vasconcelos & dos Santos, 2023). Unlike static digital tools that simply deliver information, LLMs respond dynamically to student input, allowing learners to iteratively test ideas, revise prompts, and explore alternative perspectives in real time. This responsiveness makes LLMs particularly compatible with constructionist learning environments, where knowledge emerges through experimentation, dialogue, and reflection.

This approach reflects a growing recognition that AI can support deeper learning when used within intentional pedagogical frameworks (Arar, Tlili, et al., 2025; Huang et al., 2025; Tlili et al., 2025). This pedagogical emphasis also complements recent AI leadership scholarship, which argues that ethical and effective AI integration depends not only on technical adoption but on leadership practices that foreground equity, human judgment, and educational purpose (DeMatthews et al., 2026). In the context of higher education, such leadership must extend into the design of teaching and learning processes themselves.

A constructionist approach to teaching LLMs emphasizes learning through active creation and iterative reflection, positioning students as agents in shaping their own work and thinking. When generative AI substitutes for student thinking rather than supporting it, learners risk accumulating the cognitive debt described above. In such cases, assignments may appear complete while the learner’s conceptual understanding remains unchanged. Rather than using LLMs to complete tasks, students are invited to engage in purposeful experimentation, examining how their choices influence meaning, voice, and ethical outcomes.

For example, students may generate a piece of writing using an LLM and then annotate or critique the LLM-generated contributions. Students may examine questions such as: What assumptions does the LLM make? What biases are present? How might the tool affect authorship, originality, or accuracy? This reflection process supports deeper learning, metacognition, critical judgment, and moral reasoning, core components of responsible technology use (Lee et al., 2024; Wang et al., 2025). By treating LLM outputs as a first draft rather than a finished product, students remain in control of their intellectual process, positioning themselves as active creators rather than passive consumers (Xie, 2023).

Constructionist assignments are rooted in student agency, creativity, and relevance (Kafai & Burke, 2015; Papert & Harel, 1991; Xie, 2023). LLM-supported assignments might include designing prompt strategies for writing goals, using an LLM as a writing coach, or building LLM-assisted personal statements. Students can develop and assess classroom AI use policies or collaboratively develop guidelines for permitted AI in group work. What distinguishes these assignments is not the use of technology alone, but the opportunity for students to engage in the process of building, iterating, and sharing while receiving feedback from peers and the instructor. These learning activities serve as launching pads for discussions about authorship, bias, ownership, and the broader societal implications of AI. Students can engage in a learning cycle that encourages ethical awareness and responsibility.

In constructionist learning environments, faculty shift from content experts to facilitators and co-learners, modeling curiosity, humility, and ethical discernment (Resnick, 2017; Vasconcelos & dos Santos, 2023). This pedagogical stance is especially important in the context of AI, where norms and best practices are still emerging. By engaging with LLMs alongside students (e.g., sharing examples of how they are experimenting with these tools), faculty foster a culture of transparent, dialogic learning. Faculty scaffold students’ understanding through timely feedback and co-constructed norms. A constructionist approach to teaching LLMs doesn’t simply equip students with technical skills; it invites them into the learning process and helps them develop critical dispositions they will need to navigate increasingly complex digital environments confidently, ethically, and thoughtfully. This approach also aligns with the broader potential of AI in higher education to foster integrity, innovation, and resilience within academic ecosystems (Arar, Özen, et al., 2025; Mustafa et al., 2024; Theresiawati et al., 2025). The constructionist principles highlight both the opportunities and the pedagogical risks associated with generative AI in higher education, raising important questions about how instructors can intentionally design learning environments that preserve student thinking and agency.

Challenges and Opportunities of Generative AI in the Classroom

The relevance of generative AI in higher education could not be more pressing. This section highlights how rapid industry development of educational AI tools is reshaping the pedagogical landscape and increasing the need for intentional instructional frameworks. As AI continues to evolve at a rapid pace, industries outside of education see its potential to transform learning and, in some cases, to infiltrate it in ways that bypass traditional pedagogical values. Business and tech sectors are racing to develop the next big innovation in AI for education. Major companies such as OpenAI, Google, and Anthropic are developing education-specific generative AI and forging partnerships to embed AI tools directly into classrooms and learning management systems. Generative AI tools are no longer general-purpose assistants; they are increasingly being tailored for student use, teacher support, and even automated assessment (OpenAI, 2024). As these technologies become embedded in educational infrastructures, the need for pedagogical frameworks that guide students in engaging with AI critically and intentionally becomes increasingly urgent. While these developments bring exciting possibilities, they also pose significant risks. If educators are not actively teaching students to engage with AI critically, ethically, and reflectively, we risk allowing these technologies to reshape education in ways that prioritize convenience over learning and efficiency over equity, a concern echoed in leadership scholarship calling for ethically grounded AI policy and human-centered implementation (Arar et al., 2024; Arar, Tlili, et al., 2025). Educators and academic leaders have a responsibility to model thoughtful AI use and ensure these tools enhance—rather than erode—the values and integrity of higher education. By taking a proactive, pedagogically grounded approach, educators can help ensure AI becomes a positive force in academic spaces.

Instructional Model for Human-Centered Generative AI Engagement

The instructional model presented in this study draws on these constructionist principles to structure student engagement with LLMs. To support students in becoming intentional, ethical, and effective users of LLMs in higher education, we introduce the Instructional Model for Human-Centered Generative AI Engagement (Figure 1). As a conceptual contribution, the model was developed by synthesizing established learning theory with the authors’ instructional experience and ongoing research on generative AI in higher education, rather than derived from a single empirical study. The model is organized around five recursive phases: critical and ethical awareness, prompt literacy, AI-supported learning, reflection and revision, and independent application. Each phase supports the development of student learning capacities that guide thoughtful and responsible use of LLMs. This instructional framework is designed to help faculty scaffold LLM-related learning by integrating hands-on experimentation, ethical reflection, and purposeful interaction with LLMs. This framework responds to broader calls within educational leadership scholarship for ethically grounded, human-centered, and evidence-based approaches to AI integration across educational systems (Arar, Özen, et al., 2025).

Figure 1

Drawing from critical pedagogy (Freire, 1970), constructionist learning theory (Papert & Harel, 1991), and an ethics of care that foregrounds relational and human-centered teaching (Noddings, 2003), this framework offers a human-centered and ethically grounded approach to learning with LLMs. At its core, the framework assumes that generative AI cannot replace the relational elements that sustain meaningful learning. Feedback, encouragement, empathy, and cultural responsiveness remain central to the human-centered approach that guides the instructional model. Prompt literacy (Chan, 2023; Walter, 2024) serves as the central pedagogical mechanism within the framework. Rather than teaching students simply how to use LLMs, the framework emphasizes learning through creation, iteration, and dialogue, supporting students in developing the critical consciousness, AI literacy, and ethical grounding needed to engage responsibly with AI across disciplines. This pedagogical framework also resonates with DeMatthews et al.’s (2026) argument that AI use in educational settings should be guided by leadership domains that emphasize ethical integration, instructional innovation, equity, and stakeholder engagement. Whereas their framework addresses AI leadership at the school and systems level, the present model extends that conversation into the higher education classroom by focusing on how instructors can cultivate reflective, ethical, and human-centered engagement with LLMs in daily teaching practice.

While much public discourse around AI emphasizes productivity, automation, or compliance, the Instructional Model for Human-Centered Generative AI Engagement repositions generative AI as a thinking partner that can extend and deepen learning when approached critically and reflectively. The five phases provide a process for instructors to introduce LLMs and guide students to become conscious, ethical, and effective users through a constructionist approach. Through this process, students develop learning capacities that deepen understanding, engagement, and agency.

Prompt Literacy Cycle

The Prompt Literacy Cycle (Figure 2) is the instructional tool nested within Phase Two of the Instructional Model for Human-Centered Generative AI Engagement, designed to give students a concrete, repeatable structure for developing prompt literacy. While the five-phase model guides the overall arc of LLM-integrated learning, the Prompt Literacy Cycle operates at the level of daily practice, helping students engage with LLMs through recursive experimentation, rhetorical awareness, and ethical reflection. Prompting involves technical decision-making and critical interpretation, which can be difficult to develop (Gattupalli et al., 2023; Walter, 2024). To support instructors in teaching prompt-related skills that go beyond output optimization (Arar et al., 2024; Gattupalli et al., 2023; Lee et al., 2024; Walter, 2024), the cycle offers a tangible structure that faculty can use to demystify LLMs and embed critical prompt literacy into their teaching. Instructors can introduce it through demonstrations, scaffolded assignments, or workshops, encouraging students to internalize the mechanics of prompting and the ethical and epistemological dimensions of engaging with LLMs (Chan, 2023; Gattupalli et al., 2023).

Figure 2

The Cycle guides students through a structured process that encourages experimentation, reflection, and interactive improvement in their interactions with LLMs.

1. Clarify Purpose – Students begin by defining their intention. What are they trying to explore, produce, or understand? This ensures prompts align with meaningful learning objectives.

2. Craft the Prompt – Students design their initial prompt, considering tone, scope, audience, and specificity. Instructors model how minor language changes dramatically alter generative LLM output.

3. Engage with Output – Students review the LLM-generated text critically. They annotate or evaluate its strengths and weaknesses in terms of relevance, bias, completeness, or tone. This stage builds skills in judgment and discernment.

4. Refine the Prompt – Students revise their prompt and re-engage with the LLM based on their evaluation. They explore how changes to the input affect the quality and focus of the output, promoting experimentation and iteration.

5. Reflect – Students reflect on the entire process. What did they learn about how prompts function? What assumptions or perspectives shaped the LLM’s response? How will they apply this awareness to future interactions?

Integrating the Prompt Literacy Cycle can cultivate deeper student agency, intentionality, and confidence in working with LLMs.

Teaching Applications for Generative AI–Supported Learning

Building on the Instructional Model for Human-Centered Generative AI Engagement, this section outlines teaching strategies and assignments for hands-on, exploratory engagement with LLMs. The learning activities are designed to build technical fluency with LLMs and to support ethical awareness, critical reflection, and learner agency. Rooted in the principles of constructionism, these strategies position students as co-constructors of knowledge, exploring LLMs through creation, iteration, and inquiry. Through this application, students develop practical skills and a deeper conceptual understanding. Among these activities, the Compare Prompt Literacy vs. Prompt Engineering exercise in Phase Two is particularly significant. By asking students to map the differences between technical optimization and rhetorical, ethical engagement, the activity makes the paper’s central conceptual distinction directly teachable and provides students with a concrete framework for understanding why prompt literacy matters beyond simply producing better AI outputs. Table 3 illustrates this and other classroom activities aligned with each phase of the instructional model.

Instructor Strategies to Deepen Understanding of LLM Use

Faculty mediation often determines whether AI functions as a shortcut that bypasses learning or as a cognitive partner that deepens it. Instructors play a pivotal role in shaping how students perceive and interact with LLMs. Their influence extends beyond teaching functionality: it involves modeling ethical awareness, critical thinking, and intentional engagement. In a constructionist classroom, where learners actively build their understanding through exploration and creation, instructors frame AI, not as a shortcut, but as a cognitive partner (Papert & Harel, 1991; Resnick, 2017).

One of the most effective strategies faculty can use is transparent modeling of LLM use. Rather than avoiding or obscuring AI use, instructors can openly demonstrate how they use LLMs for ideation, drafting, revision, or problem-solving. This reinforces the values of the Instructional Model for Human-Centered Generative AI Engagement, which prioritizes reflective, ethical, and dialogic interaction with LLMs. For example, instructors can craft and refine prompts, assess LLM-generated content, and iterate in real time, sharing their thought process aloud: “Here is why I phrased it this way” or “Let’s see how the AI interpreted this prompt.” These moments model intentionality, discernment, and ethical awareness in action (Chan, 2023; Walter, 2024). When AI is used in this reflective, intentional way, it supports deeper learning, critical thinking, and repeated questioning (Arar, Özen, et al., 2025; Tlili et al., 2025).

The Prompt Literacy Cycle offers a practical structure for this type of modeling. As instructors guide students through the five recursive stages: clarify, craft, engage, refine, and reflect, they model the same process in their own work. This reinforces that LLM use is not about achieving perfection or efficiency, but about engaging thoughtfully and iteratively. Demonstrating how vague prompts produce generic results and how minor revisions can yield more nuanced responses helps students see the cognitive complexity involved in effective LLM use. This repeated engagement reflects the principle of human-machine symbiosis, the collaborative relationship in which human judgment guides the process while LLM capabilities extend and deepen thinking rather than replacing it (Arar et al., 2024).

Equally important is helping students understand when, why, and whether to use AI. Instructors can frame LLMs not merely as tools for convenience but as rhetorical and ethical instruments. By prompting students to articulate the intent of their LLM use and to reflect on its impact, faculty support the development of metacognition, academic integrity, and ethical reasoning (Chan, 2023; Walter, 2024). Instructors can share their own examples, such as using LLMs to brainstorm a lesson or assignment. This helps normalize responsible experimentation and use. Research shows intentional, context-aware AI use prevents misuse and encourages meaningful engagement (Mustafa et al., 2024).

Instructors empower students by building classroom norms for LLM use together. When instructors open dialogue about appropriate usage and limitations, students are more likely to take ownership of ethical practices. Collaboratively developed guidelines foster a culture of transparency, accountability, and care (Noddings, 2003). These discussions can also be linked directly to institutional academic integrity policies, helping students understand how authorship, attribution, and responsible LLM use intersect in academic work. When norm-setting is treated as a teaching moment rather than a compliance exercise, students develop a working understanding of authorship as something they actively construct and protect. Shared norms position ethical AI engagement as a collective responsibility rather than an individual choice, aligning with broader calls for unified and ethically grounded approaches to generative AI in education (Bozkurt et al., 2024).

Finally, instructors should underscore that while generative AI offers powerful support, it cannot replace human presence in learning. Relational elements such as feedback, encouragement, empathy, and cultural responsiveness remain irreplaceable, and this is not incidental to the framework — it is central to it. The Instructional Model for Human-Centered Generative AI Engagement is human-centered in name because it is human-centered in design. When educators use AI to enhance, not replace, their instructional role, they model a balanced approach that values both innovation and humanity (Arar, Özen, et al., 2025; Bixler & Ceballos, 2025).

Instructors function as critical mediators of AI literacy and classroom culture. By teaching AI use with openness and intention, they help students develop technical skills, reflective habits, and ethical awareness. Instructors help reposition generative AI not as a threat or shortcut, but as a meaningful extension of student-centered, values-driven learning.

Conclusion

As generative AI continues to evolve from an assistive technology to increasingly autonomous systems, higher education faces not only a pedagogical challenge but a structural one. Recent systematic evaluations suggest that AI’s revolutionary potential lies in its capacity to promote active knowledge construction, but only when directed by deliberate pedagogy. The emergence of the “ghost student” and the risk of accumulating cognitive debt expose the limitations of reactive or prohibition-based approaches to AI. The response must be purposeful, proactive pedagogy: one that preserves human thinking and intellectual engagement at the center of the learning process.

In this manuscript, we introduced the Instructional Model for Human-Centered Generative AI Engagement, a constructionist framework designed to help instructors guide students in engaging with LLMs as a thinking partner rather than a substitute for learning. By emphasizing prompt literacy, reflective engagement, and iterative learning processes, the framework supports instructors in integrating LLMs into teaching while maintaining student agency and intellectual ownership.

Through transparency, norm-setting, and thoughtful integration, educators can shape a culture in which generative AI serves as a springboard for inquiry rather than a replacement for thinking. As scholars increasingly call for assessment models that foreground human presence, reflection, and dialogic engagement over measurable output alone, this framework offers a practical and pedagogically grounded response: one designed not to resist AI but to ensure it cannot replace the learner.

As a conceptual framework, the model now invites empirical evaluation, which is currently underway. Future research should explore how this framework functions across diverse educational contexts. Empirical studies could test the model’s effectiveness in supporting ethical and reflective AI engagement, examine disciplinary differences in generative AI use, and investigate how students perceive and develop prompt literacy over time. Additional research could also explore how faculty implement the framework across institutional contexts and how it intersects with evolving academic integrity policies.

These questions also connect to broader conversations in educational leadership. Scholars argue that meaningful AI integration requires ethical, equity-centered, and human-guided leadership across educational systems. While their work focuses on leadership at the school and system levels, the present study extends this conversation into higher education pedagogy by illustrating what human-centered AI engagement can look like within the classroom itself. As educational leadership scholarship emphasizes, the successful integration of AI depends not only on technological innovation but on ethical, equity-centered, and human-guided implementation across educational systems.

Data Accessibility Statement

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Sustainable Development Goals (SDGs)

This study is linked to the following SDGs: Quality Education (SDG 4).

Author Contributions (CRediT)

April Joy Miles: Conceptualization, methodology, investigation, formal analysis, visualization, writing—original draft preparation, writing—review and editing;

Paige Haber-Curran: Conceptualization, resources, writing—original draft preparation, writing—review and editing;

Khalid Arar: Conceptualization, supervision, validation, project administration, writing—review and editing; All authors have read and agreed to the published version of the manuscript.

Author Notes

Generative AI tools, including ChatGPT, Claude, Grammarly, and NotebookLM, were used during the development of this manuscript to support brainstorming, organization, language refinement, and editorial review. These tools assisted with drafting suggestions, improving clarity, and identifying areas for revision; however, all conceptualization, analysis, interpretation, and scholarly arguments were developed and evaluated by the authors. AI tools did not generate original research findings, conduct data analysis, or make intellectual contributions that would qualify for authorship. Consistent with recommendations from the Committee on Publication Ethics (COPE), the authors assume full responsibility for the accuracy, originality, and integrity of the manuscript. All AI-assisted content was critically reviewed, revised, and approved by the authors prior to submission.