AI + instructional judgment · Process case study

Using AI to design better learning—not bypass judgment.

An evidence-based look at how AI-assisted prototypes become accurate, accessible, practical learning experiences.

ContextAI Literacy, 2026
AudienceCTE students / educators
FrameworkHuman-in-the-loop review
EvidenceLive course + assets

01 / The performance challenge

Teach people how to judge, not just generate.

The Genesee County AI Literacy initiative needed a consistent, approachable learning experience spanning prompting, ethics, privacy, bias, and career applications. A central design choice was to treat AI output as a starting point to evaluate—not as an automatic answer.

02 / Design workflow

A human-reviewed AI development cycle.

01Define the decision

Identify what the learner must recognize, decide, or do in real work.

02Prototype quickly

Use AI-assisted drafting and design tools to explore multiple explanations, practice examples, and visual treatments.

03Audit the output

Apply fact-checking, privacy, bias, copyright, readability, and accessibility checks before accepting a draft.

04Make practice authentic

Use learner-facing activities, career situations, and checks for understanding—not just explanatory prose.

05Test and refine

Review content and UI, revise for student clarity, and maintain alignment between objectives and assessments.

The public course teaches a five-tip prompting framework and a three-step audit routine. Those instructional concepts also exemplify the human-judgment approach brought to AI-supported production.

03 / Artifacts

Show the work, including the review.

Three-step audit framework from the AI Literacy course
Original instructional asset from the public AI Literacy source repository.

Further course artifacts include a privacy scanner, bias awareness visuals, career-focused activities, lesson-level assessments, and educator implementation resources.

04 / Responsible design

Where human review matters.

  • Do not publish unverified facts or misleading AI-generated examples.
  • Use accessible language and alternatives to image-only explanations.
  • Protect personal information and minimize unnecessary data.
  • Separate original learner outcomes from planned enrollment targets.
  • Make materials editable and reviewable by collaborating educators.

Scope note: This is a reconstruction of the design approach reflected in the public course and its development context; it is not a claim that every draft followed an identical automated process or that a formal A/B test occurred.

Let's make learning useful.

Let's talk about building effective learning experiences for teams and organizations.

Contact Andrew ↗