Design rubric criteria and model instructions in observable, verifiable terms with citable evidence requirements¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article recommends formulating assessment criteria in observable and verifiable terms, avoiding formulations assessable only through an overall reading. Instructors should explicitly require the model to support each score with a concrete reference to the report text (citable evidence), include a structured checklist per item replicating the teaching team's checks, and ask the model to state explicitly when it cannot verify an aspect for lack of legible evidence rather than completing the response with a plausible but unsupported interpretation.

Design Implications¶

Context¶

Requirements¶

• Rubric criteria written in observable, verifiable terms; instructions requiring citable evidence for each score.

Constraints¶

• These conditions improve clarity and evidential basis but the article ties them to making instructor verification easier, not to guaranteeing correct assessment.

Target Learners¶

• higher education instructors designing AI-supported assessment

Target Learning Goals¶

• evidence-based feedback on laboratory reports

Related Strategies¶

• Use joint AI processing of report sets to identify recurring difficulties and guide collective feedback and teaching priorities
• Rubrics
• Mitigate extraction errors by requiring high-resolution PDFs, equation-editor equations, and legible graph axes, units, and legends
• Make output verification an individually accountable practice within team assignments via rubric requirements

Examples¶

-

Key Sources¶

• Marcos Abreu, Cecilia Stari, Arturo C. Martí. (2026). AI-Assisted Assessment of Experimental Physics Laboratory Reports: Potential, Limitations, and Support for Teaching Practice. arXiv. https://arxiv.org/abs/2609.22417