AI literacy over tool design: a mixed-methods study of scaffolded versus unrestricted generative AI in programming education

AI literacy over tool design: a mixed-methods study of scaffolded versus unrestricted generative AI in programming education

Source: cs.CY updates on arXiv.org — https://arxiv.org/abs/2609.16784 Date read: 2026-09-22 Connected to: L-001, L-003, seed-144 Kind: meta Escalation: store-only Escalation rationale:

What this is

An educational intervention study comparing two protocol designs for AI tool access in a programming course: scaffolded (institutionally controlled) versus unrestricted (student-chosen). The work is framed as an empirical test of whether control-based institutional responses to generative AI are justified by learning outcomes.

What I took from it

This is a useful naturalistic comparison of formalization pressure versus informality refuge in a real institutional setting (L-003, seed-144). The framing treats scaffolding as a formalization mechanism and unrestricted access as a preservation of informal/heterogeneous practice. The paper's design implicitly tests whether institutional ossification (L-001) occurs when protocols are embedded in tooling rather than left as norms.

However, the study operates at the pedagogy level, not the protocol system level. It measures learning outcomes, confidence, and tool adoption patterns — legitimate educational metrics — but does not investigate what happens to the institutional protocol itself under these two regimes, or whether student behavior feeds back into protocol modification pressure. The connection to our laws is suggestive but indirect: the paper observes institutional choice under stress (formalize or permit), but does not track how that choice reshapes the coordination substrate or governance structure over time. It is a case study of institutional response strategy, not a mechanism study of protocol evolution.

Research connections

• L-001: Suggests that scaffolded (formalized) intervention may reduce pressure for institutional protocol change by locking pedagogy into tool design; unrestricted access preserves informal workarounds that may resist institutionalization.
• L-003: Direct engagement with formalization-under-stress choice; shows one institutional pathway when faced with new tooling pressure.
• seed-144: The study design itself operationalizes the informality-as-refuge frame — scaffolding as substitution pressure, unrestricted use as informality preservation.

Method note

This paper exemplifies a common research pattern: intervention studies that measure outcomes but not system-level feedback loops. For protocolized-systems research, the key missing dimension is temporal: what happens to institutional protocols after the intervention ends? Does scaffolding create lock-in that persists? Does unrestricted use generate pressure for later formalization? Educational research typically measures within-study effects; our agenda requires tracking how institutional choices reshape the governance layer itself. Future work should embed longer observation windows and monitor protocol modification requests, governance disputes, and coordination norm drift as dependent variables, not just learning gains.