Retrofitting an AIAS level to an unchanged assessment task creates a validity risk in which the assessment measures use of the scale rather than intended learning outcomes¶

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

Subclaims¶

q2 i? When the AIAS level becomes the object of design rather than a means of aligning GenAI use with learning outcomes, assessment may no longer generate the intended evidence of learning. → Perkins 2026

Evidence¶

Perkins 2026¶

Perkins, M., Postma, D., Roe, J., Sisay, S., & Holdcroft, C. (2026). 'A bit of chaos and madness': The AI Assessment Scale and the work of assessment reform. Preprint. https://arxiv.org/abs/2606.26729 doi:10.48550/arxiv.2606.26729

q2 · i? · qualitative · r3

Theme 6 (Reframing Pedagogy) of the thematic analysis. One participant reported carrying assessments over with little redesign while assigning a level, while others used the scale to redesign tasks toward authentic research evidence; the authors say this empirically confirms an anticipated failure mode.

"when the AIAS level is treated as the object of design rather than as a means of aligning GenAI use with learning outcomes, the assessment may no longer generate the evidence of learning it was intended to capture."

Discussion¶

Related Claims¶

• Existing GenAI assessment frameworks (traffic light, two-lane, AIAS) remain broad, lack discipline-specific applicability, and are unvalidated in authentic settings — related
• Unclear policy, workload pressure, and lack of mechanisms to monitor student AI use undermined staff confidence and perceived fairness during AIAS implementation — related
• Academic staff experienced the AIAS as a shared language that legitimised GenAI use in teaching and assessment — related
• The article claims LLM-assisted workflows weaken the validity of output-based evaluation in hiring, education, and certification — a broader claim this one bears on