Back to Course

Responsible AI for Healthy and Thriving Learners — Principles, Practice and Policy

0% Complete
0/0 Steps
  1. Policy, Principles and Practical Implementation
    5 Topics
  2. Foundations: Key Definitions and How to Use This Course
    3 Topics
  3. Responsible AI Innovation for Young People and Educators
    6 Topics
  4. Navigating the Boundary: Educational AI vs. Health Services
    5 Topics
  5. AI’s Impacts on Young People’s Well‑Being
    5 Topics
Lesson Progress
0% Complete

Watercolor header showing a small, diverse group of educators and young people — different ages, races, genders, visible disabilities (wheelchair, hearing aid) — gathered around a table with an inclusive-design checklist (Yes/Partial/No), sticky notes, pens, and a tablet displaying a simple "AI disclosure" icon. A nearby whiteboard shows a 2×2 priority matrix and accessibility icons (captions, headphones, keyboard). Warm, collaborative, non-blaming mood; soft muted palette, gentle brush strokes, paper texture, and clear negative space for a title.

A hands‑on checklist activity to evaluate a tool or lesson for inclusivity, accessibility and youth participation.

This activity helps educators, designers and policy makers systematically check whether a learning tool or lesson is inclusive, accessible and genuinely centered on young people’s participation and well‑being — especially when the tool uses or interacts with AI.


Why this activity?

You can’t fix what you don’t measure. This quick, practical checklist gives teams (and young people) a shared language for spotting problems, prioritizing fixes, and documenting decisions. Use it with a prototype, an existing product, or a lesson plan.


Learning objectives

  • Assess a tool or lesson across inclusion, accessibility, and youth participation criteria.
  • Identify concrete, prioritized changes to improve safety and fairness.
  • Practice doing evaluations in partnership with young people.

Time, group size & materials

  • Time: 45–90 minutes (depends on depth)
    • Quick scan: 30–45 minutes
    • Full co‑design review with youth: 60–90 minutes
  • Group size: 2–6 people per review group (include 1–2 young people when possible)
  • Materials: checklist printouts or digital copy, the tool or lesson materials, sticky notes, whiteboard or shared doc, device to demo the tool

Before you start

  1. Pick the deliverable to review (lesson plan, app, prototype, chatbot script).
  2. Gather relevant artifacts: privacy policy, learning outcomes, screenshots, data flow diagrams, transcripts.
  3. Invite at least one young person representative (age-appropriate, paid/honored for their time where possible) to join the review or run the activity with a youth co‑design session.

How to run the activity (step‑by‑step)

  1. Quick orientation (5–10 min)
    • Explain purpose and process. Set a positive, non‑blaming tone: we’re evaluating, not policing.
  2. Individual scan (10–20 min)
    • Each participant (including youth) uses the checklist to score the tool privately and write comments.
  3. Small group discussion (20–30 min)
    • Compare scores, note agreements/disagreements. Capture evidence and examples.
  4. Prioritization (10–15 min)
    • Use a simple priority matrix (High/Low effort × High/Low impact) to choose 2–4 immediate actions.
  5. Action planning (10–15 min)
    • Assign owners, timelines, and next steps. Document where youth input is required for the fix.
  6. Optional deeper testing
    • If serious issues show up, schedule follow-up co‑design or usability testing with a more diverse group of young people.

The Inclusive Design Checklist

For each item mark: Yes / Partial / No — and add notes/evidence + suggested fix.

Section A — Representation & Relevance

  • A1. Learning goals and content reflect diverse identities (race, gender, sexuality, ability, socioeconomic background). Why it matters: content that assumes a single “norm” excludes many learners.
  • A2. Examples and imagery include young people like those in your audience. (Not tokenistic.) Why it matters: representation builds belonging.
  • A3. Content avoids stereotypes and pathologizing language. Why it matters: language shapes mindsets.

Section B — Accessibility (WCAG mindset)

  • B1. Text content follows plain language best practices for the target age. Why it matters: comprehension is core to inclusion.
  • B2. Visuals have alt text or accessible descriptions. Why it matters: users with visual impairments need content alternatives.
  • B3. Videos include captions and transcripts. Why it matters: supports D/deaf learners and those who prefer reading.
  • B4. Interactive elements are keyboard‑navigable and screen‑reader friendly. Why it matters: ensures usable for many disability profiles.
  • B5. Color contrast and font sizes meet readable standards; no information conveyed by color alone. Why it matters: visual accessibility.
  • B6. Adjustable pacing and multiple ways to engage with content (text, audio, visuals, practice). Why it matters: supports neurodiversity and different learning styles.

Section C — Language, Comprehension & Cultural Sensitivity

  • C1. Reading level matches the intended age group. Why it matters: prevents confusion and disengagement.
  • C2. Content is localized (not just translated). Examples and idioms make sense locally. Why it matters: cultural relevance matters.
  • C3. Terms around sex, gender, and relationships are inclusive, accurate and age‑appropriate. Why it matters: supports healthy, respectful learning.

Section D — Youth Participation & Agency

  • D1. Young people were consulted or co‑designed this tool/lesson. Why it matters: increases relevance and trust.
  • D2. The design gives learners meaningful choices (topics, pace, privacy options). Why it matters: promotes autonomy.
  • D3. There are clear mechanisms for learners to provide feedback and see changes. Why it matters: sustains continuous improvement.
  • D4. Consent and assent for data collection are designed for young audiences (clear, layered, age‑appropriate). Why it matters: ethical and legal obligations.

Section E — Safety, Emotional Well‑Being & Safeguarding

  • E1. The tool includes safe exits and on‑ramps to human support when needed. Why it matters: prevents harm escalation.
  • E2. Content is trauma‑informed and avoids retraumatization triggers; trigger warnings are used thoughtfully. Why it matters: protects vulnerable learners.
  • E3. Moderation and reporting processes are clear and youth‑friendly. Why it matters: enables safe participation.
  • E4. For sexuality education: content promotes consent, respect and accurate information; avoids shaming. Why it matters: central to learner well‑being.

Section F — Power, Privacy & Data Practices (AI-relevant)

  • F1. The tool discloses if and how AI is used, in plain language for young people and caregivers. Why it matters: transparency builds understanding and trust.
  • F2. Data collection is minimized to what’s necessary; retention policies are clear. Why it matters: reduces privacy risk.
  • F3. Opt‑out options are available and easy to use (for both accounts and data collection). Why it matters: respects autonomy.
  • F4. Sensitive attributes (health, sexual orientation, gender identity) are not inferred or used without explicit, ethical justification and consent. Why it matters: avoids harm and unjust profiling.
  • F5. There is human oversight for decisions that significantly affect learners (grading, disciplinary actions, personalized recommendations). Why it matters: algorithmic decisions should not replace human judgment.
  • F6. Bias audit or fairness checks have been performed; known limitations are documented. Why it matters: reduces discriminatory outcomes.

Section G — Evaluation & Iteration

  • G1. There is an evaluation plan that includes diverse youth voices (surveys, interviews, observations). Why it matters: measures real-world impact.
  • G2. Version history and change rationale are documented (why changes were made). Why it matters: accountability and institutional memory.
  • G3. There is a roadmap for addressing identified issues, with timelines and responsibilities. Why it matters: moves from audit to action.

Scoring & Interpretation (quick method)

  • For each item: Yes = 2, Partial = 1, No = 0.
  • Sum total and compute percent of maximum possible.
    • 80–100%: Good baseline — still iterate with youth.
    • 50–79%: Needs work — prioritize quick fixes and safety concerns.
    • <50%: High risk — pause deployment until major issues are resolved.

Focus first on items that affect safety, privacy and rights (sections E and F) even if the overall score looks okay.


Prioritization and action planning (simple template)

After scoring, fill this:

  • Issue: e.g., “No captions on videos”
  • Priority: High / Medium / Low (consider legal/safety impact)
  • Proposed fix: e.g., “Add captions and transcripts”
  • Owner: name or role
  • Timeline: due date
  • Youth input needed?: Yes/No; how?

Pro tip: Use a quick 2×2 grid (Impact vs. Effort) to pick 2 high‑impact/low‑effort wins to do within one sprint.


Facilitation tips when working with young people

  • Compensate or otherwise acknowledge youth contributions.
  • Use plain language and give examples to explain checklist items.
  • Make sessions short, interactive and safe (breakout groups, polls, anonymous feedback options).
  • Be trauma‑informed: warn about sensitive topics and allow opt‑outs without pressure.
  • Include diverse young people (different ages, abilities, languages, cultural backgrounds).
  • If the tool handles sexual health or personal data, include safeguarding protocols and adult support on call.

Quick example (sample completed items — hypothetical AI tutoring app)

  • A1 Representation: Partial — examples limited to urban school contexts. Fix: add rural scenarios.
  • B3 Video captions: No — videos lack captions. Fix: add captions & transcripts (High priority).
  • D1 Youth co‑design: No — no youth input. Fix: recruit teen advisory group (Medium/High).
  • F1 AI disclosure: Partial — mentions “AI” in long T&Cs. Fix: add layered, plain‑language banner at sign-up explaining AI features (High).
  • E1 Human support: Yes — live tutor escalation available. Note: great, ensure tutors are trained for sensitive subjects.

Total score: calculate to prioritize captioning, AI transparency, and youth advisory group as immediate actions.


Adaptations

  • Remote review: use shared doc + screen share; collect anonymous input via forms.
  • Large organizations: run parallel small groups, then synthesize common findings.
  • Policy makers: use the checklist as a baseline for procurement requirements and require vendors to provide evidence (accessibility reports, bias audits, data minimization statements).

Closing notes

  • This checklist is a living tool. Revisit it regularly and refine it with young people.
  • Don’t treat “Yes” as perfect — document evidence and keep improving.
  • For legal or complex data questions, consult your institution’s legal and safeguarding teams.

Ready-to-use checklist table (copyable)

| Item code | Item | Yes / Partial / No | Notes / Evidence | Suggested fix |
|—|—:|—:|—|—|
| A1 | Diverse representation in content | | | |
| A2 | Inclusive imagery/examples | | | |
| B1 | Plain language | | | |
| B2 | Alt text for images | | | |
| B3 | Captions/transcripts | | | |
| B4 | Keyboard/screen-reader friendly | | | |
| C1 | Reading level appropriate | | | |
| D1 | Youth consulted/co‑designed | | | |
| D2 | Learner choices available | | | |
| E1 | Human support/escalation | | | |
| E2 | Trauma‑informed content | | | |
| F1 | AI disclosure in plain language | | | |
| F4 | Sensitive inference avoided | | | |
| G1 | Evaluation plan with youth | | | |

Use this template to run your first review. Take notes, act on the high‑impact fixes, and then bring a more diverse group of young people back for round two. Small, consistent improvements make tech and lessons far safer and more inclusive for learners.