Integrate co-design and continuous feedback loops into AI evaluation processes¶

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

Description¶

The report recommends that states integrate co-design with feedback loops—defined as structures that are purposeful, multi-directional, iterative, and transformative—into evaluation across all implementation stages. Co-design communities via AI task forces, working groups, and dedicated committees help include multiple contributing groups' voices and create continuous cycles of improvement responsive to diverse community input, scaling and evolving innovations.

Design Implications¶

Context¶

Requirements¶

• Dedicated groups such as AI task forces or working groups responsible for developing or overseeing AI policy and guidance
• Integration of student, teacher, parent, and community voices in development, review, and oversight

Constraints¶

Target Learners¶

• State education leaders
• Educators
• Students
• Parents and caregivers

Target Learning Goals¶

• Responsible, evidence-based adoption of AI-enabled technologies in education

Affordances¶

• Three Stage Ai Evaluation Stages Framework

Claims¶

• Eleven of 33 state guidance documents mention co-design or feedback loop structures, and 27 include student, teacher, parent, and community voices [~W]

Related Strategies¶

• Establish AI-specific accountability systems with iterative playbooks, family guides, and continuous evaluation cycles
• Build state AI evaluation on the ESSA evidence framework and integrate results into the existing education evidence base
• Ensure communication and co-development with school leaders throughout edtech process rollout
• Use iterative design with short feedback loops and multiple user-feedback methods during product development
• Center co-design with teachers and students throughout AI tool R&D, using human-in-the-loop workflows as a primary risk mitigation
• Five-step iterative co-design process for developing edtech evaluation criteria with learners and vendors
• Establish continuous improvement cycles that monitor and evaluate POG implementation
• Use a design thinking process with Empathize, Define, Ideate, Prototype, and Test stages to reimagine school safety
• Use generative AI to make assessment systems more instructionally relevant, not to reify problematic practices
• Involve domain experts throughout the AI safety evaluation pipeline, especially for defining unsafe content
• Use the DOT Framework's Empathize and Test/Feedback stages as practical checklists, and build professional learning communities around shared prompting and critique practices
• Staged institutional implementation: audit interactions, map to teaming levels, build AI literacy, co-design, and evaluate iteratively
• Integrate GPT models into LA-supported feedback systems with an iterative instructor-student validation loop to improve reliability

Examples¶

-

Key Sources¶

• Ruiz, P., Prado, Y., Roschelle, J., & Fusco, J. (2025). What States Say About Evaluating AI in Education: Reviewing Guidance From 32 States and Puerto Rico. Digital Promise. https://doi.org/10.51388/20.500.12265/280