# Part 0. How AI Works (/how-ai-works/overview) --- type: Document title: "Part 0. How AI Works" description: "What Part 0 teaches about how AI works and the skills to work with it, and who can skip it." status: stable order: 10 ksor: owner: team:panaversity audience: [ public ] approval: by: process:panaversity at: 2026-10-12T16:49:58Z part: "0" expert_status: null objectives: [] word_budget: null prerequisites: [ "F2" ] build_step: null artifact: null field_guides: [] delta_entries: [] concepts: [] chapters: [ "A", "B", "C", "D" ] running_project: "The AP Worker at Brightline Wholesale Supply" alignment: anthropic: "Supporting coverage of CCAO-F Domains 1, 2, 3, 4 and 6, and the first introduction to CCDV-F D5.S1" openai_competencies: [ OAI.1.1, OAI.1.2, OAI.1.5 ] next_exam: "PCAO-F, after Part II" last_verified: null generated: at: 2026-10-12T16:49:58Z by: process:panaversity trust_tier: unverified build_id: sha256:3051b1b7d623cf82892eef720f02c742ae47da9c0c7e09e60c3000529a25d63f dirty: true ksor_version: 0.0.60 --- > [!NOTE] > **You are here: Part 0. How AI Works** > > | Journey | Where you are | > | --- | --- | > | Capability | You may never have used an AI assistant, or you have used one without knowing how it works. This part explains the machine and names the four competencies you need, so the rest of the book makes sense. | > | Certification | No credential yet. This part prepares the ground for PCAO-F, the Panaversity exam you take after Part II. | > | Economic | You can explain to a colleague why AI sometimes sounds right and is wrong, and what a fluent user does about it. Everyone who earns from AI work needs that first. | > > - Portfolio artifacts: four short pieces of work, one from each chapter. Two of them carry into Part II. > - Who reads it: every route, but it is optional. If you pass its skip quiz, you can go straight to Part I. > - Before you start: the pages before Part 0, especially [How to Use This Book](../how-to-use-this-book.md). > - What comes next: Part I shows how the machine that predicts text became a worker you manage. Part II goes deeper into each of the four competencies. ## What this part is about This book starts from one idea: > We have moved from the chatbot paradigm to the AI Worker paradigm. The winners will be the people, companies and countries that learn to manage, build, govern, deploy and export AI work. In plain words, AI used to answer questions in a chat. Now it can also take on whole tasks, like a worker. Part 0 explains the machine behind that change, so the rest of the book makes sense. Part 0 explains how AI works. It is written for everyone, including readers who have never used an AI assistant. It has four chapters, labeled A to D. Chapters A to C explain the machine behind the AI text box. Chapter D covers the human side. It names four competencies, the skills you need to work with AI effectively, efficiently, ethically and safely. Parts I and II then teach those competencies in depth. Part 0 never teaches the procedures of briefing, review or governance. A brief is the written instructions for one task. Chapters 5, 6 and 10, in Part II, teach those procedures. ## Who can skip it Part 0 is optional on every route. It adds up to 2 hours. Take its skip quiz first. The quiz has about 20 questions on the four chapters, and a score of 80 percent or more is a pass. If you pass, go straight to Part I. The chapters and the skip quiz arrive in a later release, so for now, start with [Part I](../ai-worker-paradigm/overview.md). ## The four chapters Each chapter has a hands-on lab and one piece of work for you to keep. The labs for Chapters A, C and D use Brightline Wholesale Supply, the company in the book's running example. Brightline is a fictional distributor in Columbus, Ohio. Every week, its office receives bills from vendors, called invoices. Someone has to record them, check them and prepare them for payment. That work is accounts payable (AP). | Chapter | What it teaches | What you make | | --- | --- | --- | | A. What AI Actually Is | What happens behind the AI text box, in nine ideas. For example, AI predicts text instead of looking facts up, and nothing in it checks the truth. | AI field notes, a one-page sheet with one example from your own tests for each idea | | B. Your First Hour with Claude and ChatGPT | The setup Part I assumes. You open accounts on Claude and ChatGPT, set the privacy controls before any real data goes in, and tour the other controls. | A setup checklist for both Claude and ChatGPT | | C. Markdown: The Language AI Reads and Writes | Markdown is a simple way to mark headings, lists and tables in plain text. You learn why structure helps AI, and when to ask it for a page, a document or a table. | A first structured markdown page, made from Brightline's messy AP notes | | D. AI Fluency: The 4Ds | The four competencies of the AI Fluency Framework, by Rick Dakan and Joseph Feller with Anthropic. They are Delegation, Description, Discernment and Diligence, the 4Ds for short. | A 4D plan for one Brightline AP task | ## How Part 0 connects to Parts I and II Part 0 gives you the ideas and the first pieces of work that Parts I and II build on. - Chapter A's last group of ideas shows how AI can use tools and work in steps. Chapter 1, which opens Part I, starts from that idea. - Chapter C's markdown page becomes the first draft of Brightline's KSoR in Chapter 8. KSoR, the Knowledge System of Record, is how a company keeps the knowledge it has approved. - Chapter D's 4D plan is the input to Chapter 5's brief and Chapter 6's Review Contract. A Review Contract is how you will know the work is right, agreed before the work starts. - Chapter D ends with a map. For each of the 4Ds, it names the chapters in Parts I and II that teach it in more depth.