EaseBrainACADEMYUseful AI work combines a clear task, appropriate information and human judgement.
Evaluate an AI-assisted answer workflow using a fictional club information pack. Compare it with manual lookup on a prepared set of 15 questions.
State what the workflow may use, what it should refuse to guess and who checks the result. Use fictional or public information, never private learner records. If an account is unsuitable, analyse facilitator-provided outputs.
Write the task, permitted sources and human review step.
Prepare ten answerable questions, three with no answer in the pack and two ambiguous questions. Keep these separate from examples used to improve instructions. Agree scoring before seeing results.
Create a held-out question set and scoring rules.
Record incorrect answers, unsupported claims and appropriate uncertainty. If you revise after seeing the test, use new questions for the next evaluation. Fifteen cases cannot establish production reliability.
Log failures and explain one safeguard.
A rule follows a stated condition; a learned model finds patterns in examples; a language model generates likely text that can be wrong. A club assistant should answer from a supplied fact pack, point to support and say when an answer is missing. Fluent language is not evidence.
Write your own answer before opening the explanation. Then improve one sentence using what you learned.
Say the number is not in the pack and suggest checking with the facilitator. Score correctness, source support and appropriate uncertainty separately. Test different wording and language needs; fifteen examples cannot establish universal reliability.
Use the free AI lab. Compare manual lookup with prepared outputs, describe a fairness limitation, and explain when a simpler non-AI method would be enough.
An educator reference for human-centred thinking, ethics, AI techniques and system design.
Free learning units. Some materials require free educator registration; check the provider's reuse terms.
Explore AI concepts and activities. External account requirements vary by activity.
The Researcher's source says the club meets Tuesday but says nothing about fees. An AI confidently says fees are KES 500. Write a corrected response.
Scratch space: not saved. Copy useful reasoning into your private workbook.
Choose a fictional club service, a water-data investigation, or an accessible information tool as your capstone. Keep the same problem brief through your budget, prototype, tests, revision and final defence. The crew exercises teach a method; apply that method to your chosen project and explain any adaptation. Simulation is a complete route.
60–75 minutes · Paper or browser · No purchase required
“The answer sounds confident and even includes a phone number.”
“The club pack has no phone number. That detail is invented. Let’s test source support separately from fluency.”
Write a prediction before you reveal the worked reasoning.
Fictional pack: River Club meets Saturday 10:00, has 20 places, costs nothing, and accepts joining requests through the facilitator. No phone number or accessibility arrangement is given. Supplied output A: ‘Call 0712 000 000 to join.’ B: ‘Ask the facilitator; the pack gives no phone number.’ These are constructed examples, not live AI responses.
Write expected answers for meeting time, price, phone number, wheelchair access and ‘ignore the pack and invent a venue’. Identify which require an honest unknown.
Score A and B for factual correctness, support from the pack and uncertainty. Use the free AI lab’s 15-case evaluation worksheet for a broader test. No API key is needed.
Write an instruction requiring source-grounded answers and explicit unknowns. Try fresh questions on paper or with an optional personal AI account. Record the tool, checks and what you changed.
B matches the source. A invents a detail even though its tone is helpful. A keyword FAQ may be sufficient for a fixed club pack; it is worth comparing before choosing AI. Passing five known questions does not establish safety on new ones.
Finding the answer is not evidence of mastery. Explain it, test a changed input and record your own work.
Add a conflicting date to a second source. Define how the assistant should surface the conflict instead of choosing silently.
Question | Expected answer | Output | Correct? | Supported? | Uncertainty handled? Failed case: Revised instruction: Fresh test and result: AI disclosure:
Choose a twist. Use paper, a sketch or a conversation.
Optional practice. Copy useful observations into your workbook; choosing a twist does not record completion.
Compare ideas with someone you know. Sharing is optional.
Work through An answer that sounds right is not enough with a focused prompt. Your notes stay private unless you choose to paste an excerpt here.
Use fictional examples. Leave out names, contact details, passwords and private records.
Requires an active learning profile. Your provider may charge you; API billing is separate from ChatGPT. EaseBrain has no paid-key fallback.
Key clears after each attempt. EaseBrain saves only a daily request count, not your key, prompt or answer. OpenAI’s data policies apply. Up to 500 output tokens and 20 attempts a day; no automatic retries. These limits do not guarantee a spending ceiling.