Fostering Self-Directed Growth with Generative AI: Toward a New Learning Analytics Framework

Synthesis: Mao (2025) proposes a conceptual framework that integrates Generative AI (GAI) and learning analytics to cultivate Self-Directed Growth — a competency enabling learners to iteratively drive their own developmental pathways across contexts. At its core is the Aspire to Potentials for Learners (A2PL) model, which reconceptualizes learner aspirations, complex thinking, and summative Self-Assessment within GAI-supported environments. The paper argues that sustainable learner Learner Agency and educational equity depend less on access than on building the Capacity to Aspire and clear developmental pathways, positioning GAI as a non-prescriptive collaborative scaffold rather than a content provider.

Key Findings

• Self-Directed Growth as a distinct competency. The paper defines it as a learner's capacity to drive continuous and sustainable self-directed learning (SDL) cycles across diverse tasks and contexts, positioning the learner as initiator, executor, assessor, and primary agent of their own journey. It is framed as an advanced extension of self-directed learning oriented toward self-actualization and an internal compass rather than predefined task outcomes.
• The A2PL model reconfigures three pillars of SDL. Drawing on Morris's (2019) foundations, the framework centers the Capacity to Aspire (humanistic philosophy), Complex Thinking (Constructivism epistemology), and Self-Assessment (pragmatic philosophy) around Self-Directed Growth, operating across three interacting levels: Person, Task, and Person x Task. This is grounded in learning-theory traditions connecting aspiration, complex reasoning, and reflexive self-evaluation.
• GAI as a non-prescriptive collaborative scaffold. The paper argues GAI should support guided discovery rather than deliver direct answers or disclose SDL rubrics, using diagnostic analytics (scoring summative self-assessment frameworks against an Aspire to Potential Scoring Rubric) and interactive analytics (sequenced, thought-provoking, calibrated prompts) to avoid reinforcing over-reliance on AI.
• A methodological critique of task-level agency measurement. The paper criticizes existing SDL/SRL studies that rely on post-task questionnaires and task-specific Self-Efficacy as indicators of agency, arguing this risks circular reasoning and neglects context, strategies, and feedback integration. It calls for dynamic, learner-driven learning analytics that capture analytical cognition of SDL across cycles.
• Equity through the Capacity to Aspire, not just access. Building on Appadurai's (2004) capacity to aspire and UNESCO's SDG 4 framing, the paper contends that educational equity requires learners — especially disadvantaged populations — to envision and pursue diverse, personalized pathways aligned with their own experiences, rather than relying on standardization or information access alone.
• Conceptual status explicitly acknowledged. The study is theoretical and non-empirical, with acknowledged limits: under-theorized human–AI psychological dynamics, a need for clearer operational definitions of Self-Directed Growth, and the absence of critical theories of power, access, and technological determinism.

What this means for practice

• Learning analytics designers. Instrument GAI as a non-prescriptive scaffold: score learner self-assessment against the Aspire to Potential Scoring Rubric as diagnostic analytics and sequence calibrated, thought-provoking prompts as interactive analytics, withholding the SDL rubric so the learner — not the model — holds the judgment.
• Learning analytics designers. Capture analytical cognition across successive SDL cycles instead of relying on post-task questionnaires and task-specific Self-Efficacy, which the paper criticizes as circular indicators of Learner Agency.
• Designers. Build for guided discovery rather than answer delivery, so that analytics-driven prompting does not reproduce the over-reliance it is meant to counter.
• Designers. Treat equity as a design goal about the Capacity to Aspire — helping learners envision and pursue personalized pathways — rather than as a matter of tool access or standardized content.

Limitations

• Purely conceptual: the A2PL model and the GAI–LA integration are theoretical proposals with no empirical validation, pilot, or sample.
• The framework assumes learners will engage with AI-supported feedback reflectively and strategically; variability in intrinsic motivation, metacognitive maturity, and socio-cultural positioning is acknowledged but not accounted for in the model.
• Human–AI psychological dynamics remain under-theorized — perceived AI authority, externalization of judgment, and erosion of internal self-regulation are named as open questions rather than addressed.
• Self-Directed Growth still lacks clear operational definitions distinguishing iterative self-directed cycles from long-term aspirational trajectories, and the framework draws only on humanistic, constructivist, and pragmatic traditions, omitting critical theories of power, access, and technological determinism.

Citation

Mao, Q. (2025). Fostering Self-Directed Growth with Generative AI: Toward a New Learning Analytics Framework. [cs.CY].