What Is Generative AI?
What Is Generative AI? keeps Gen AI grounded instead of magical thinking. Phase focus: Mindset & Mechanics. Anchor idea: Generative AI predicts plausible next pieces—tokens—rather than …
GENERATIVE AI
Learn Gen AI with judgement—not hype
Six playful arcs move from mechanics to responsible mastery. Each lesson blends plain-language explanations with three tap-only checks on this page. Answer them correctly, unlock the next topic—same rhythm as our Python track.
6 arcs
24 lessons
Micro-check gates
Common myths—straight answers
Quick reality checks that match what search engines call FAQ markup—same wording lives in structured data for clarity.
Usually no. Many setups answer from model weights plus whatever context, tools, or documents you provide. Live browsing or retrieval only happens when the product explicitly enables those features.
No. Fluent tone is not evidence of truth. Models can "hallucinate"—invent plausible facts, links, or citations. Verify anything important against primary sources.
It means the model produced confident-looking content that is wrong or unsupported—often because it predicted plausible tokens rather than retrieving verified facts.
Avoid it. Treat prompts as sensitive notes: redact secrets, use placeholders, and follow your school or workplace policy. Logs and integrations can expose what you type.
Not automatically—it depends on rules and how you use it. Coaching yourself with practice questions is different from submitting AI-written work as entirely your own.
Not always. Clear goals, audience, and format beat sheer length. Noise and contradictions in giant prompts can confuse the model.
It retrieves relevant snippets from documents first, then generates an answer grounded in those snippets—still imperfect, but closer to cited evidence than guessing from memory alone.
Sometimes—that family of risks is called prompt injection. Keep trusted instructions separate from random web dumps or suspicious attachments.
How this course feels different
- Short reads first—mental models you can reuse in exams and internships.
- Each lesson ends with three multiple-choice checks written for that topic—graded instantly in your browser.
- Truth-telling beats buzzwords: hallucinations, privacy, bias, and disclosure show up early.
- Unlock progression keeps the roadmap lightweight—you always know what is next.
Lesson deck
Progress: 0/24
How Large Language Models Think (Simply)
Lesson spotlight: How Large Language Models Think (Simply). Phase context: Mindset & Mechanics. How experts picture it: An LLM scores likely continuations from patterns seen during …
Tokens and the Context Window
Welcome to Tokens and the Context Window. You are in the Mindset & Mechanics arc—bite-sized, playful, but serious about accuracy. Big picture: Text splits into tokens; …
Training vs Inference (Student-Friendly)
Training vs Inference (Student-Friendly) keeps Gen AI grounded instead of magical thinking. Phase focus: Mindset & Mechanics. Anchor idea: Training builds weights once; inference runs those …
Hallucinations and Confident Wrong Answers
Lesson spotlight: Hallucinations and Confident Wrong Answers. Phase context: Mindset & Mechanics. How experts picture it: Fluent ≠ accurate; models can invent plausible citations or figures. …
Prompt Clarity: Task, Audience, Format
Welcome to Prompt Clarity: Task, Audience, Format. You are in the Prompt Craft arc—bite-sized, playful, but serious about accuracy. Big picture: Explicit structure beats vibes—state goal, …
System Roles and Tone Without Drama
System Roles and Tone Without Drama keeps Gen AI grounded instead of magical thinking. Phase focus: Prompt Craft. Anchor idea: Roles anchor behaviour—teaching assistant vs critic …
Few-Shot Examples Inside a Prompt
Lesson spotlight: Few-Shot Examples Inside a Prompt. Phase context: Prompt Craft. How experts picture it: Two crisp examples teach format faster than long prose instructions. Keep …
Chain-of-Thought Light (Show Your Steps)
Welcome to Chain-of-Thought Light (Show Your Steps). You are in the Prompt Craft arc—bite-sized, playful, but serious about accuracy. Big picture: Asking for intermediate reasoning can …
Iteration: Ask, Critique, Improve
Iteration: Ask, Critique, Improve keeps Gen AI grounded instead of magical thinking. Phase focus: Prompt Craft. Anchor idea: Treat prompting like editing loops—each pass sharpens structure. …
Privacy: What Not to Paste into AI Tools
Lesson spotlight: Privacy: What Not to Paste into AI Tools. Phase context: Safety & Integrity. How experts picture it: Assume prompts may be logged—protect IDs, keys, …
Bias, Blind Spots, and Respectful Use
Welcome to Bias, Blind Spots, and Respectful Use. You are in the Safety & Integrity arc—bite-sized, playful, but serious about accuracy. Big picture: Training mirrors messy …
AI for Learning vs Shortcut Cheating
AI for Learning vs Shortcut Cheating keeps Gen AI grounded instead of magical thinking. Phase focus: Safety & Integrity. Anchor idea: Explain-it-to-me loops cement understanding; copy-paste …
Strong Study Workflow With AI Tutoring
Lesson spotlight: Strong Study Workflow With AI Tutoring. Phase context: Study Workflows. How experts picture it: Alternate retrieval practice with explanations—AI coaches spaced repetition. Keep it …
Summaries That Stay Faithful
Welcome to Summaries That Stay Faithful. You are in the Study Workflows arc—bite-sized, playful, but serious about accuracy. Big picture: Ask for quotes + citations from …
Debugging Your Thinking (Not Just Homework)
Debugging Your Thinking (Not Just Homework) keeps Gen AI grounded instead of magical thinking. Phase focus: Study Workflows. Anchor idea: Expose mistaken assumptions; AI can stress-test …
Draft Smarter: Outlines First, Polish Later
Lesson spotlight: Draft Smarter: Outlines First, Polish Later. Phase context: Study Workflows. How experts picture it: Structured outlines curb rambling and hallucinated citations. Keep it conversational—pair …
Retrieval-Augmented Generation (RAG) Simply Explained
Welcome to Retrieval-Augmented Generation (RAG) Simply Explained. You are in the Systems & Multimodal arc—bite-sized, playful, but serious about accuracy. Big picture: RAG retrieves snippets then …
Agents, Plugins, and Tool Use (Big Picture)
Agents, Plugins, and Tool Use (Big Picture) keeps Gen AI grounded instead of magical thinking. Phase focus: Systems & Multimodal. Anchor idea: Agents loop plan-act-observe with …
Multimodal AI: Text, Images, and Beyond
Lesson spotlight: Multimodal AI: Text, Images, and Beyond. Phase context: Systems & Multimodal. How experts picture it: Vision/audio adds cues—still probabilistic and culturally biased. Keep it …
Prompt Injection Awareness (Everyday Level)
Welcome to Prompt Injection Awareness (Everyday Level). You are in the Responsible Mastery arc—bite-sized, playful, but serious about accuracy. Big picture: Malicious text can steer assistants—treat …
Image Generation and Prompt Basics for Arts
Image Generation and Prompt Basics for Arts keeps Gen AI grounded instead of magical thinking. Phase focus: Responsible Mastery. Anchor idea: Diffusion models blend noise reduction …
Disclosure, Citations, and Fact Checking
Lesson spotlight: Disclosure, Citations, and Fact Checking. Phase context: Responsible Mastery. How experts picture it: Label AI assistance; cite primary sources you verified. Keep it conversational—pair …
Building Your Personal AI Study Playbook
Welcome to Building Your Personal AI Study Playbook. You are in the Responsible Mastery arc—bite-sized, playful, but serious about accuracy. Big picture: Codify habits—prompt templates, review …