Class 7 · CBSE AI · Strand D — The Architect's Capstone
How to run an AI ethics review — fairness, privacy and harm
Who benefits, who could be harmed, what data is used: the four lenses of an AI ethics review. For Class 7.
What this concept actually says
- An ethics review examines who benefits, who might be harmed, and what data is collected or used by the proposed AI
- A defensible ethics position can be explained to a sceptic — it identifies risks and specifies what safeguards address each risk
- Bias, privacy, consent, and transparency are the four core lenses of an AI ethics review
An analogy your child will recognise
Medicine prescription
A doctor who prescribes medicine knows the benefit but also writes the side effects on the prescription. They don't say 'let's hope it works out.' They name exactly what could go wrong and what the patient should watch for. An AI ethics review is the same discipline — name the side effects, not just the cure.
Water from a new borewell
Before a new borewell is used in a village, someone tests the water for impurities. The question isn't 'does it look clean?' — it's 'is it actually safe for the specific people who will drink it?' An ethics review asks the same rigorous question about your AI before it reaches real users.
Common misconceptions to watch for
- Ethics reviews are only needed for big, powerful AI systems — even a small school project that processes personal data needs one
- If your intentions are good, the ethics are fine — harm can come from well-intentioned systems; intent and impact are different things
Key facts in one breath
- The four core lenses of AI ethics are: fairness (does it disadvantage a group?), privacy (what data is collected and who sees it?), transparency (can users understand what the AI is doing?), and accountability (who is responsible when it goes wrong?)
- A written ethics review creates a record — it allows reviewers, users, and future builders to see what was considered and what was not
- The EU AI Act (2024) requires high-risk AI systems to have documented risk assessments — ethics reviews are not academic exercises
- Children's data has special legal protections in many countries; any AI handling data of under-18s must meet a higher bar for consent and storage
How Dhee Learning teaches this — the 3-stage question loop
Every Dhee Learning session for this concept follows three stages. We share the questions Dhee actually asks, so you can hear what a session sounds like.
Stage 1 — Surface
Your AI helps teachers identify students who are 'likely to fail' so the teacher can give them extra help. Who does this help? Can you think of a way it might hurt someone?
Rote answer
"Child only sees the benefit (teacher helps struggling students) without surfacing any potential harm"
Understood
"Child raises at least one: being labelled 'likely to fail' might affect how the teacher treats the child, the prediction might be wrong (false positive), the data used might be biased against certain groups, or the child/parent might not have consented to being predicted on"
Stage 2 — Reasoning
Your AI collects voice recordings of children to work better over time. Write three specific questions you would put in an ethics review document before you're allowed to launch this feature.
Follow-up Dhee may use: If you can't answer one of those questions yet, does that mean you can't launch, or just that you need to work harder to find the answer?
Stage 3 — Application
Write a 150-word ethics summary for your own capstone project. It must cover: (1) who benefits and how, (2) one specific risk of harm, (3) one concrete safeguard you have built in or plan to build in. Make it defensible — someone sceptical of AI should be able to read it and see you've thought carefully.
Misconception Dhee watches for: Child writes a benefits-only summary with a single token sentence about 'making sure it's fair' — no specific risks, no specific safeguards, and no testable commitments
Related concepts
Want your child to actually understand this?
Dhee turns this concept into a 15-minute spoken session — asking, listening, and probing — so your child builds the idea themselves.
Frequently asked questions
What is ethics review — written, defensible — explained for kids? +
Who benefits, who could be harmed, what data is used: the four lenses of an AI ethics review. For Class 7.
What's the most common mistake children make about this concept? +
Ethics reviews are only needed for big, powerful AI systems — even a small school project that processes personal data needs one
How does Dhee Learning teach this in a Class 7 session? +
Dhee opens with a question — for example: "Your AI helps teachers identify students who are 'likely to fail' so the teacher can give them extra help. Who does this help? Can you think of a way it might hurt someone?" — listens to your child's answer, then probes the reasoning behind it. The session ends when the child can apply the idea to a brand-new situation, not just recall it.