Within this LMS-integrated design, personalised GPT use fell into four recurring cognitive partnership roles: conceptual tutor, co-designer, reflective mirror, and cognitive challenger¶

Claim · All claims
Evidence · 1 study · design r3 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Four typologies of GPT use emerged from the combined dataset, representing distinct cognitive partnership modes across participants. → Jason Zagami 2026

Evidence¶

Jason Zagami 2026¶

Jason Zagami. (2026). Generative AI as a cognitive partner: personalised GPTs and differentiated learning in undergraduate teacher education. Education Tech Research Dev. https://doi.org/10.1007/s11423-026-10746-4

q2 · i? · design · r3

Thematic analysis of logs, journals, artefacts and focus groups from 14 pre-service teachers in a 12-week design-based study identified the four roles; Table 3 profiles each by primary function and learner profile, e.g. the conceptual tutor "Simplifying explanations, providing analogies" for novice learners.

"From the combined dataset, four typologies of GPT use emerged, summarised in Table 2 and 3. These represent distinct cognitive partnership modes observable across participants."

Discussion¶

Related Claims¶

• Students' main AI challenges are content reliability and depth, and most expect AI to serve as tutor, creative partner, and critical thinking partner rather than mere content generator — related
• Students showed increasing metacognitive regulation across the semester, moving from copying GPT suggestions to predicting and challenging its responses — related
• Client interviews suggest a recurring shift from treating AI as a passive answer machine (\"Oracle\") to engaging it as a collaborative tool under human direction (\"Process Partner\") — related
• Lower-experience students used GPTs as conceptual tutors while technically stronger students used them as professional collaborators, consistent with adaptive scaffolding — related
• Students perceived high learning impact from the personalised GPT agents, with improved conceptual understanding rated highest (mean 4.6, SD 0.5) — related