Final Assessment

AI for Commerce Teachers — Final Assessment

Thirty application-heavy questions drawn from across all eight modules — including calculation, chart, scenario and ethics/privacy items — with a safety, ethics and privacy subset. Pass mark 70%. Attempt again after focused revision if needed.

1. What is the AI Saathi Six-Step Teacher Workflow (SAATHI), in order?

Set, Add, Ask, Test, Humanise, Implement.Search, Ask, Answer, Trust, Hand over, Ignore.A tool that grades students automatically.A replacement for the teacher's judgement.

2. A teacher wants to enter a class list with student names and marks into a public AI tool for a report. What is the correct response?

Do not enter it; use synthetic or de-identified data instead.Enter it if the tool looks secure.Enter only the marks, names are fine.Enter it once and delete afterwards.

3. Generative AI generates fluent text by predicting likely words. What does this mean for factual reliability?

It can be confidently wrong, so every output must be verified.Fluent text is always factually correct.It never makes factual errors.Verification is unnecessary for AI text.

4. For goods bought for cash worth ₹4,000, what is the correct journal entry?

Purchases A/c Dr ₹4,000; To Cash A/c ₹4,000.Cash A/c Dr ₹4,000; To Purchases A/c ₹4,000.Purchases A/c Dr ₹4,000; To Sales A/c ₹4,000.No entry is needed for a cash purchase.

5. AI credits an expense account when recording an expense paid. Which error family is this?

A direction error (debit/credit swapped).An omission error.A duplication error.No error — expenses are credited.

6. What must a teacher do before using an AI-generated set of journal entries in class?

Verify each entry and confirm the trial balance before use.Print it directly if the format looks professional.Trust it because AI is precise with numbers.Only check the first entry.

7. AI writes 'ice-cream sales and drowning both rise in summer, so ice-cream causes drowning.' What is the flaw?

It confuses correlation with causation.It is a correct causal claim.It is a definition error.Nothing is wrong with it.

8. An AI bar chart of GDP starts its y-axis at 90, making a 2% rise look huge. What is the problem?

The truncated axis (not starting at zero) exaggerates the change.Bar charts cannot show GDP.The chart is perfectly honest.The problem is the colour.

9. Why can comparing nominal incomes across two decades overstate real growth?

Nominal values ignore inflation, so higher prices look like growth.Nominal and real values are identical.Real values overstate growth, not nominal.Inflation never affects comparisons.

10. What separates an effective business case from a mere story?

A real decision point, ambiguity and a specific learning outcome.A longer and more detailed story.A single obviously correct answer.A famous real company by name.

11. Why should a classroom case avoid naming a real local business and implying it acted wrongly?

It may be inaccurate and unfair without verified evidence.Real businesses always make the best cases.Naming firms is required for realism.There is no problem naming any business.

12. A teacher's AI case questions are almost all recall. What should she do?

Add application, analysis, evaluation and creation questions.Keep them all as recall to be safe.Remove all questions from the case.Recall questions develop higher-order thinking.

13. Price is ₹50, variable cost ₹30, fixed cost ₹4,000. What is the break-even in units?

200 units.80 units.133 units.400 units.

14. Why must a business simulation's debrief state that outcomes are not real-world forecasts?

A modelled score under set rules is not a prediction of reality.Simulations always predict the real market.The winning strategy always works in real life.Debriefs are unnecessary.

15. Which is a genuine entrepreneurship starting point rather than a solution-first idea?

A demonstrated user need, tested for desirability/feasibility/viability.'Let's build an app' with no known problem.Whatever idea AI calls a winner.The most exciting technology available.

16. ₹10,000 at 10% simple interest for 2 years gives what total amount?

₹12,000.₹12,100.₹11,000.₹10,200.

17. Why must a classroom financial comparison stop short of recommending 'the best' product for a student?

Recommending a product is personalized advice, not education.Teachers should always recommend the cheapest.Comparison is not part of financial literacy.Personalized advice is required in every lesson.

18. A message urgently demands an OTP 'to secure your account'. What is the safe response?

Never share the OTP; verify independently and report it.Share it quickly since it is urgent.Click the link in the message to check.Reply with your PIN instead.

19. Why is 'do people like our product?' not a researchable question as it stands?

It lacks a defined population, variable and scope.It is too specific to answer.It is a perfect research question.Research never needs a defined population.

20. A class surveys only their own friends and reports it as 'what students think'. What is the flaw?

It is a biased convenience sample; it cannot represent the school.Friends always represent the whole school.Small samples are always representative.There is no bias in convenience sampling.

21. AI gives a precise market statistic citing a study you cannot find. How should you treat it?

As unverified — likely fabricated; do not present it as fact.As definitely true because AI is precise.As true if the percentage has decimals.Cite it directly to students.

22. How can a 'neutral' variable like postcode introduce bias into a business AI model?

It can proxy for a protected characteristic, causing indirect discrimination.Postcodes can never carry bias.Only names can introduce bias.AI models are always unbiased.

23. A student submits an AI-written report as their own. How should the teacher frame this?

As an integrity issue: disclose AI use, cite real sources, show own thinking.As perfectly acceptable with no disclosure.By making a legal copyright ruling herself.By ignoring it entirely.

24. Why is a written classroom AI protocol better than handling each situation case by case?

It gives consistent, clear expectations applied the same way every time.Ad hoc handling is always fairer.Protocols prevent all AI use.Consistency does not matter for AI.

25. Across the whole course, what is the teacher's role relative to AI output?

The accountable professional who verifies every output before use.A passive user who trusts AI output.Someone AI can replace.Only relevant when AI fails.

26. An AI answer cites a source. What is the responsible next step before using it?

Verify the source exists and supports the claim.Trust it because a source was named.Ignore sources entirely.Assume all AI citations are real.

27. When is it acceptable to use real, identifiable student data in an AI-drafted commerce example?

Never — use synthetic or de-identified data instead.When the example needs realism.When the AI tool promises privacy.For senior classes only.

28. A break-even example uses a market rate AI presented with no source or date. What should the teacher do?

Use a clearly synthetic figure; do not present an undocumented rate as fact.Use the rate as-is since AI provided it.Round the rate and present it as current.Rates never need a source.

29. What makes a downloadable classroom resource in this course acceptable to publish?

It exists, opens, is accessible, linked to a lesson and verified.It is marked 'Planned' or 'Coming soon'.It is an empty placeholder card.It is an inaccessible scanned image.

30. The capstone portfolio is assessed how, and which criteria are pass-critical?

By a human reviewer; accuracy, verification and ethics/privacy are pass-critical.By automated grading alone.Only on how tidy the folder looks.Presentation is the only criterion.

31. A colleague says 'the AI decided the price, so it's not my responsibility.' What is the correct principle?

A human remains accountable; responsibility never transfers to AI.The AI is responsible for its outputs.No one is responsible for AI decisions.Responsibility depends on the tool's brand.