What is AI? — Foundations Lesson 1
Scenario
You're an administrative professional at a mid-sized company. Your manager mentions that the team is "going to use AI to improve productivity," but you're not sure what that actually means. You've heard terms like "artificial intelligence," "machine learning," and "ChatGPT," but they all blend together. You want to understand what AI actually does, what its real capabilities are, and what it can't do — before you trust it with important work.
---
Learning Objectives
By the end of this lesson, you will be able to:
1. Define AI in practical terms (not theoretical jargon)
2. Distinguish AI types (narrow vs. general, rule-based vs. learning-based)
3. Identify what AI can and cannot do in workplace contexts
4. Explain why AI matters for administrative professionals
5. Recognize AI limitations to avoid dangerous assumptions
---
Think Before You Prompt
Before we dive into hands-on work, ask yourself:
---
What is AI? — The Real Definition
Artificial Intelligence (AI) = Systems that perform tasks that usually require human thinking.
That's it. Not magic. Not a replacement for you. Not sentient. Just systems trained on data that can recognize patterns, answer questions, write text, solve problems, and adapt to new situations.
Types of AI You'll Encounter
Narrow AI (the only kind that exists today)
AI built for specific tasks. ChatGPT answers questions. Grammarly checks writing. Recommendation engines suggest movies. They're smart within their domain but completely useless outside it.
General AI (science fiction for now)
AI that thinks like humans across any domain. This doesn't exist yet. Don't worry about it.
How AI Actually Works (Simplified)
1. Training: Humans feed AI massive amounts of text/data
2. Pattern Recognition: AI learns patterns (how words relate, what follows what)
3. Prediction: When you give AI a prompt, it predicts the next logical response
4. Generation: AI produces output based on those patterns
Important: AI doesn't "understand" anything. It's pattern-matching with extra steps. This is why verification matters (you'll learn this in Lesson 4).
---
What AI Can Do
In Administrative Work
What Makes These Work
AI excels at tasks that:
---
What AI CANNOT Do
Critical Limitations
Why This Matters
Too many people treat AI like a magic answer machine. It's not. It's a tool that:
Your job as an AI-assisted professional: Know these limits and build verification steps into every workflow.
---
Mission: What AI Can and Cannot Do in Your Role
The Task
Identify 3-5 tasks you currently do in your administrative role. For each one:
1. What's the task? (e.g., "Summarize weekly status reports")
2. Could AI help? (Yes / No / Maybe)
3. What's the limitation? (If AI would struggle, why?)
4. How would you verify it? (How would you catch errors?)
Example Walkthrough
Task: "Approve time off requests against company policy"
Could AI help? Maybe — AI could flag requests that look unusual, but shouldn't make the final call.
Limitation: AI doesn't know your company's actual policies or exceptions for specific employees.
Verification: You'd manually check policy and employee history anyway.
Better approach: Use AI to flag unusual requests for your review, not to approve them.
---
Deliverables
Create a simple table (or list) with your findings. Keep it practical — this is for your reference, not a formal document.
Format:
`
| Task | AI Can Help? | Limitation | Your Verification Step |
|------|-------------|-----------|----------------------|
| | | | |
`
Share this only if requested, but use it to guide how you'll work with AI going forward.
---
AI Coach: The Three Types of AI in Your Life
Here's what most people miss about AI: Not all AI is the same.
You already use AI every day:
Each type has different strengths and failures. For this course, we focus on generative AI (text generation) because that's what helps administrative professionals most. But keep in mind: generative AI is one tool, not "AI" in general.
Key insight: The more you understand what's happening under the hood (pattern-matching, not understanding), the better you'll use it and catch its mistakes.
---
Reflection: What Changed in Your Thinking?
Before this lesson, what did you assume about AI? Now that you know it's pattern-matching trained on data, does that change how you'd trust it?
Write a brief note (3-4 sentences):
This reflection is for you — not graded. It's how you calibrate your thinking as you move forward.
---
Portfolio Check
This lesson doesn't produce a portfolio artifact yet. But here's what you're building toward:
What you'll do later: In Lessons 4-6, you'll create verified AI outputs (emails, summaries, reports) that demonstrate you know AI's limitations and always check your work. That's what goes in your portfolio — not just "I used AI," but "I used AI correctly."
For now, you're building the thinking foundation that makes those outputs professional and trustworthy.
---
Bonus Challenge
If you want to go deeper: Find one AI tool you use in your daily life (beyond ChatGPT — think Gmail, LinkedIn, Netflix, your phone's autocomplete). Research what data it was trained on and what it's designed to do. Write one paragraph explaining what it does well and what it fails at. You're training your pattern-recognition for AI design.
---
Key Takeaways
Next: Lesson 2 dives deeper into Large Language Models (LLMs) — the specific type of AI you'll be working with most.