Learner Guide

What NTPU AI Learn is

NTPU AI Learn is a learning platform made up of multiple courses. The course currently available is AI 101: Claude AI Workforce Fundamentals.

Getting started

1. 1Open the course
   Go to AI 101 from “Courses” and read the course overview first
2. 2Get a session code
   The tests, Use Case, and Exit Ticket need a session code from your instructor
3. 3Follow the path
   Learn → Practice → Assess → Apply

Browsing courses

The Courses page lists every course. Each card shows the course code, duration, level, and availability.

• Browse courses

AI 101 at a glance

AI 101 takes about 6 hours and suits complete beginners with no programming background. It includes two learning modules, four practice tasks, pre- and post-tests, and a personal Use Case.

• AI 101 course page

Learning modules

Each module is made of short sections, and the table of contents lets you jump around. The “Check yourself” questions only help you confirm understanding — they are not recorded.

• Module 1
• Module 2

Practice workflow

The practice page offers four tasks (Chat, Prompt, Cowork, Code) and the files they need. Download or preview a file first, then follow the steps.

• AI 101 Practice

Pre-test and post-test

Enter the session code and a learner code of your choice to begin. Use the same learner code for both tests. Answers are scored on the server, and you see your result after submitting.

• About the assessment

Use Case Canvas

Record one task that recurs in your work. Apart from the work name and the help you want from AI, most fields are optional.

• About applying

Exit Ticket

A short wrap-up before the course ends, used to plan follow-up department interviews.

Switching language

Use “中文 / EN” in the header. You stay on the same page, and your module and section position is kept.

Technical requirements

• A recent version of Chrome, Edge, Safari, or Firefox
• An internet-connected computer — phones and tablets work for reading, but a computer is recommended for practice
• An account that can use Claude for the practice tasks (as described by your course)

Accessibility basics

The site supports keyboard navigation, clear focus indicators, a skip-to-content link, and respects the system “reduce motion” setting. If something is not usable for you, tell your instructor.

Privacy and data use

The platform collects only the minimum needed to run the course and uses a learner code instead of your real identity. Test answers, Use Cases, and Exit Tickets are visible only to authorized teachers and administrators.

Frequently asked questions

1. 1What if I do not have a session code?
   Ask your instructor. Reading the modules and practice needs no code; only the tests and forms do.
2. 2Can I use different learner codes for the pre- and post-test?
   We recommend against it. The same code lets the results be matched.
3. 3Can I edit something after submitting?
   Not yourself. If a correction is needed, contact your instructor.

Quick reference

The material below is for review after class (formerly the “Learner Handbook”): core concepts, tool selection, prompt template, document analysis, review, and Use Case.

Getting help

During a course, ask your instructor first. If you are studying on your own, go back to the relevant module or the quick reference in this guide.

Core Concepts

• Generative AI: produces new content from your input
• Agentic AI: plans and executes multi-step tasks
• Chat / Cowork / Code: three primary work modes
• Human–AI–Human: a human defines the task, AI executes, a human reviews and stays accountable

Tool Selection

1. 1Single-turn, easy-to-state task
   → Chat
2. 2Multi-step, multi-source task
   → Cowork
3. 3Recurring, well-defined task
   → Code / automation
4. 4High-risk, high-impact task
   → Whichever mode, add stronger human review

T-C-R-F Template

• Task: the concrete, checkable job
• Context: background, purpose, constraints
• Reference: source material or examples
• Format: expected structure, length, tone

Six-Step Document Analysis

1. 1Understand
   What the document is about
2. 2Organize
   Break it into issues, owners, timeline
3. 3Analyze
   Find patterns, risks, relationships
4. 4Question
   Identify information gaps
5. 5Produce
   Turn it into a summary or recommendation
6. 6Verify
   Check key facts and numbers against the source

• Useful follow-up: "Which passage supports this conclusion?"
• Useful follow-up: "Is there any conflicting information?"
• Useful follow-up: "What here is fact versus your inference?"

Review

• Fact: explicitly stated in the document
• Analysis: inference built on facts
• Unknown: information the document does not provide, needing follow-up

• AI Review Checklist: can the conclusion be traced back?
• AI Review Checklist: do the numbers match the source?
• AI Review Checklist: any unverified inferences?
• AI Review Checklist: does it match the original task goal?

1. 1Can this data leave internal systems?
   Data-safety question 1/4
2. 2Does it contain customer PII or confidential material?
   Data-safety question 2/4
3. 3Who can see this conversation?
   Data-safety question 3/4
4. 4What's the cost of getting it wrong?
   Data-safety question 4/4

Cowork

• Delegate Task Checklist: is the scope explicit?
• Delegate Task Checklist: is the delivery format clear?
• Delegate Task Checklist: is the needed reference material provided?

• Pre-execution Planning: have AI state its plan first
• Pre-execution Planning: confirm before it continues
• Pre-execution Planning: break complex tasks into checkable stages

Code

A one-off task rarely justifies the cost of automation; recurring, well-defined tasks are the ones worth scripting.

Use Case

The Use Case Canvas records: current process, frequency, time spent, pain points, the AI help you want, suggested mode, consequence if wrong, and human-review method.

• Value check: how often does this happen, and how much time does it cost?
• Risk check: how serious is the consequence if wrong, and is the data sensitive?

Post-Class Prompt Cards