Pusaram AI Academy
Pusaram sites

Lesson O01 · AI Awareness Weekend

What AI can help with and where judgment stays with you

Draft lesson

Draft lesson. Published by {PUBLISHER}. Complete text, quiz and task, but not yet tried with learners or reviewed by an independent specialist. Examples use synthetic facts. What the status labels mean.

Your outcome

Given an everyday task, distinguish a fixed rule, a prediction, a generated draft, and a decision that needs human responsibility. No programming is required.

Begin with an ordinary situation

You are helping organize a community workshop. You need to add up costs, estimate attendance, write an invitation, and decide whether the venue is suitable. All four involve information, but they are different kinds of work. Using one AI chat for everything can conceal important differences.

A spreadsheet can calculate a total using an explicit formula. A predictive model can estimate attendance from past examples. A generative system can propose wording. A person still needs to check the venue's actual availability and decide whether it meets the group's needs.

A useful mental model

Think of an AI assistant as a capable draft partner whose work needs inspection. This analogy helps with review, but has limits: a model does not have the same lived experience, accountability, or understanding of your situation as a human colleague.

Artificial intelligence is a broad term for systems performing tasks associated with capabilities such as recognizing patterns, predicting outcomes, or generating content. Machine learning uses patterns learned from data. Generative AI creates outputs such as text, images, or audio. A product may combine a model with search, calculators, and other tools; do not assume those tools were used merely because the answer looks detailed.

Why there are different approaches

AI development includes rule-based approaches, statistical learning, neural methods, and large general-purpose models. These approaches coexist. A clear rule remains useful when the required behavior can be specified precisely. Learning from examples becomes useful when patterns are difficult to express as a complete rule. Generative systems offer flexible drafts, but flexibility also makes their outputs harder to guarantee.

You do not need to memorize a timeline today. The important historical lesson is that newer methods have expanded the available options, rather than eliminating the need for older ones.

Worked example: a workshop plan

Synthetic facts: the room holds 20 people; 16 people have confirmed; each refreshment pack costs ₹50; the refreshment budget is ₹1,000.

Task A: calculate the cost of 16 packs. Use arithmetic: 16 × 50 = ₹800. Check with a calculator. Task B: draft an invitation. AI may help with tone and structure. Task C: predict the final turnout. The supplied facts do not establish how many additional people will attend. Task D: approve the venue. A person must consider access, safety, availability, and the actual booking conditions.

An illustrative weak AI response says: “Twenty people will attend, so the venue is perfect.” It has turned a capacity into an attendance prediction and made a suitability judgment without evidence. A stronger response separates facts from unknowns: “Sixteen people are confirmed. The room has four places remaining. Final attendance and venue suitability still need checking.”

TaskSuitable methodWhat can be checkedMissing informationResponsible person
Calculate the refreshment costCalculator or written arithmetic16 × ₹50 = ₹800Nothing missing for this stepThe organizer
Draft an invitationAI drafting is a good fitWording against the supplied factsCalendar date and booking detailsThe organizer reviews and sends
Estimate final attendancePrediction from past examplesCompare with real attendance historyNo history is supplied. Room capacity (20) is not a forecastThe organizer decides what to plan for
Approve the venueA human decisionBooking confirmation, access, safetyNone of these is suppliedThe organizer
Figure 1. Four tasks, four kinds of help
Text equivalent

Four tasks from the community workshop, each with a suitable method, something to check, what is missing, and who is responsible. Calculate the refreshment cost: calculator or written arithmetic; check 16 × ₹50 = ₹800; missing: nothing missing for this step; responsible: the organizer. Draft an invitation: ai drafting is a good fit; check Wording against the supplied facts; missing: calendar date and booking details; responsible: the organizer reviews and sends. Estimate final attendance: prediction from past examples; check Compare with real attendance history; missing: no history is supplied. room capacity (20) is not a forecast; responsible: the organizer decides what to plan for. Approve the venue: a human decision; check Booking confirmation, access, safety; missing: none of these is supplied; responsible: the organizer.

When AI may add little value

A calculator is usually clearer for this small multiplication. A fixed checklist may be sufficient for checking whether required form fields are present. Choose AI where its flexibility helps, such as rewriting an invitation for a specific audience, rather than because it is available.

How to choose a method for a taskA decision flow: define the task; if the rule is explicit use a rule or calculator, otherwise ask whether a draft is useful and generate one or seek evidence or human expertise; then check facts and constraints; use the checked result only if there is enough evidence and authority, otherwise seek evidence or human expertise.YesNoNoYesYesNoDefine the taskIs the ruleexplicit?Use a ruleor calculatorIs a draftuseful?Seek evidenceor humanexpertiseGenerate a draftCheck facts andconstraintsEnough evidenceand authority?Use the checked result
Figure 2. Choosing a method for a task Original Academy diagram (specification REF-D01).
Text equivalent
  1. Define the task.
  2. Ask: is the rule explicit? If yes, use a rule or calculator, then go to step 5.
  3. If not, ask: is a draft useful? If no, seek evidence or human expertise.
  4. If a draft is useful, generate a draft.
  5. Check facts and constraints.
  6. Ask: is there enough evidence and authority? If yes, use the checked result. If no, seek evidence or human expertise.

Explicit rules can use deterministic tools; a generated draft always needs checks. An action still needs sufficient evidence and appropriate authority.

Practice and transfer

Use task O01-T01 in the learner workbook. Choose the community case or substitute a study group, family event, volunteer activity, or workplace meeting. Keep the task type the same. Identify one fact, one generated suggestion, one unknown, and one decision owner. Your occupation does not limit which example you may choose.

Try a new case: a student wants a study timetable and asks whether following it guarantees an exam result. A timetable can be drafted; an outcome cannot be guaranteed from the timetable alone. Explain what additional evidence would be needed and which choices remain the student's responsibility.

Common mistake and correction

Mistake: “If the answer is fluent and specific, the system must know.” Correction: inspect what the claim is based on. A detailed sentence may still be unsupported. Ask what information is missing and whether a simpler tool would be easier to verify.

Complete quiz O01, then record one appropriate AI use and one task you will check another way. Next: write a request that makes success easier to judge.

References and context

This unit's examples are synthetic teaching examples, not recorded model tests. Educational framing is informed by UNESCO's human-centred AI competency approach: https://www.unesco.org/en/articles/ai-competency-framework-students. Review date: 17 November 2026.

Check your understanding

This is a self-study check, not a graded test. Choose one answer for each question, then check. Explanations appear after you check, and you can try again. The answers are in this page, so it cannot prove what you know.

Question 1 of 4 Which task is most directly suited to generative drafting?
Question 2 of 4 The room holds twenty people. What does this establish?
Question 3 of 4 Why might a fixed rule be better than an AI model for checking required fields?
Question 4 of 4 A fluent answer contains an unsupported attendance estimate. What should you do?

Try it: Choose the right kind of help

About 15 minutes. Uses the practice packet (synthetic facts only).

  1. Read Practice Packet A.
  2. Classify calculation, invitation drafting, turnout prediction, and venue approval.
  3. For each, name a suitable tool or reviewer and one limitation.
  4. Explain one missing fact.

You produce: Four-row task comparison; Two-sentence explanation.

Open the full task, rubric, hints and sample answer

Selected references and optional next reading

These are supporting choices, not extra required homework for the four-hour workshop. The lesson and supplied exercise remain sufficient for this introductory outcome.

Core reading

  • Generative AI for Beginners Microsoft · repository · Public source; check repository license and individual assetsChoose relevant introductory lessons; coding/API portions are optional for awareness learners.Try it: Explain the idea and apply it to the supplied synthetic example.About this reference (G01)

Optional

Use the selected concept to explain or check your workbook task. Technical code and paid books are optional; read only the introductory passages that match the focus.

Links go to the owners' own sites. Linking to a book or repository is not permission to copy it, and some resources cost money; the lesson does not require them. Reviewed 2026-11-17. See all references.