Passive engagement within correctly classified Aid tasks is a detection-resistant mis-skilling pathway requiring expert epistemic auditing¶

Claim · All claims
Evidence · 1 study (2 entries) · theoretical r2 · q1
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q1 i? A learner who correctly identifies an Aid task but engages passively (reading without evaluating, accepting without comparing) internalizes plausible-sounding AI errors, constituting a failure of epistemic activity rather than subzone selection. → Tsim 2026 q1 i? Passive engagement is indistinguishable from active engagement within a correctly identified Aid task, so supervision requires subzone re-identification from Aid to Non-Negotiable with human experts as epistemic auditors. → Tsim 2026 (2)

Evidence¶

Tsim 2026¶

Tsim, F., Gutoreva, A., Weiss, A., & Dubosh, N. (2026). AI as Teammate: Rethinking Task Distribution in Medical Training. Correspondence: [email protected]. https://arxiv.org/abs/2608.28373 doi:10.48550/arxiv.2608.28373

q1 · i? · theoretical · r2

Theoretical argument in Section 3.3.2.3 describing an underappreciated mis-skilling form involving engagement quality rather than subzone selection, where wrong clinical structures form through the interaction rather than despite it.

"AI-generated content is processed and potentially internalized, but the incomplete knowledge base means plausible-sounding errors pass undetected, and the cognitive interaction that makes Aid-subzone engagement developmentally beneficial fails to instantiate."

Tsim 2026 (2)¶

Tsim, F., Gutoreva, A., Weiss, A., & Dubosh, N. (2026). AI as Teammate: Rethinking Task Distribution in Medical Training. Correspondence: [email protected]. https://arxiv.org/abs/2608.28373. doi:10.48550/arxiv.2608.28373

q1 · i? · theoretical · r2

Theoretical argument posing a practical supervision challenge: because passive and active engagement look identical, the proposed solution has experts audit how the learner engaged with the mode, not which mode was selected, aligning with Vygotsky's ZPD.

"Within a correctly identified Aid task, a learner passively engaging with AI is indistinguishable from the one exercising actively."

Discussion¶

Related Claims¶

• AI output errors are frequent: 44.5% fabricated references in drafted research proposals and 5–13% unsafe chatbot medical answers — related
• Learner AI over-reliance is reframed as misclassification: a failure of real-time metacognitive evaluation with two distinct pathways — related
• Task-level account of upskilling and the Triad of Skill Failure (deskilling, never-skilling, mis-skilling) with differing empirical support — a broader claim this one bears on
• Constructive learning beats active and passive learning — related
• AI in education poses challenges across three interrelated domains: knowledge itself, the process of knowing, and the socio-environmental impact of knowledge production — a broader claim this one bears on
• Three recurring AI failure modes arose in this project-based learning context: plausible-but-incorrect code, missing specialized knowledge, and limited long-term context — related
• Engagement mode is orthogonal to instructional task: any task can be engaged at any of the four modes — related