Design verification and mutual-evaluation assessments for fairness and feasibility: alternative response modes, equitable AI access, selective documentation, and planned workload¶

Strategy · All strategies
Evidence · 1 claim (1 mixed) · 1 study (1 qualitative), q2 · 0 of 1 report an effect size · 1 claim rests on one study

Description¶

The article recommends concrete design actions for fair, feasible implementation: offer alternative response modes that preserve the capability and assistance conditions being assessed; address AI access inequity by providing a common tool, supplying the same generated output for critique, or permitting equivalent non-AI pathways; keep documentation selective and proportionate; and plan workload concretely. The author urges that "an accommodation is not an evidentiary concession when it removes a barrier unrelated to the construct."

Design Implications¶

Context¶

Requirements¶

• Students should know what is collected, who can access it, how long it is retained, and how it can be used
• Where sampling is used, selection should follow a disclosed neutral rule

Constraints¶

• Alternatives need examination to ensure they demand comparable disciplinary reasoning; a response mode producing systematic disadvantage unrelated to the construct would require revision

Target Learners¶

• higher education students, including those with accommodations and varying access to AI tools

Target Learning Goals¶

• fair demonstration of the assessed capability regardless of response mode or tool access

Claims¶

• Unclear policy, workload pressure, and lack of mechanisms to monitor student AI use undermined staff confidence and perceived fairness during AIAS implementation [~W]

Related Strategies¶

• Use this Toolkit: Mitigating Racial Bias in Edtech
• Mitigating Racial Bias in Edtech Products
• AI in Education Toolkit for Racial Equity
• Implement a coordinated set of safeguards: data-protection policies, fairness-oriented algorithms, assessment redesign, teacher professional development, subsidised tools, and infrastructure investment
• Treat pathway legitimacy as a design target: ensure alternative routes are perceived as legitimate, usable, and socially recognized

Examples¶

-

Key Sources¶

• William J. Rothwell. (2026). Assessing Learning When AI Can Produce the Product: An Integrated Framework for Cognitive Ownership and Mutual Evaluation. https://orcid.org/0000-0002-4503-1957