How People Should Talk to AI (Without Losing Themselves)

AI, Humans, and the Speed of Change — Part 2 of 4

Where We Are in the Conversation

This is Part 2 of a four-part series on AI, people, and the speed of change.

In Part 1, we slowed things down and talked plainly about what AI actually is, and what it isn't. Not magic. Not consciousness. A very powerful system trained to recognize patterns and predict what comes next. Useful, yes. But still a tool.

That grounding matters. Because when people misunderstand AI, they tend to swing to one of two extremes: blind trust or total avoidance. Neither one ends well.

So now comes the practical question:

How should normal people actually talk to AI?

Not like engineers. Not like prompt wizards. And definitely not like you're trying to impress anyone.

This part is about interacting with AI in a way that's clear, natural, and sustainable, without turning yourself into a machine or feeling like you're "doing it wrong."

You Don't Need to Talk Like a Machine

A lot of people get stuck before they even start.

They assume there's a "right way" to talk to AI. Special syntax. Perfect prompts. Some secret language they missed. That pressure alone is enough to make people shut down.

Here's the simple truth:

You don't need to talk like a machine. You need to talk like yourself.

AI doesn't require fancy commands. It works best when it has context. And context comes from natural language: how you already explain things, ask questions, and think out loud.

This isn't about being casual or sloppy. It's about being clear.

Clear beats clever every time.

Should You Say "Please" and "Thank You"?

This one comes up more than you'd think. And there are two camps.

Camp 1 says: be yourself. If you're a polite person, talk to AI the way you talk to anyone. It helps AI understand your tone and write more like you.

Camp 2 says: cut the pleasantries. AI isn't a person. "Please" and "thank you" don't change the output. You're just burning tokens.

Here's the thing, both camps are partially right. And neither one is the whole story.

The real question isn't whether politeness helps the AI. It's whether forcing yourself to talk like a robot helps you.

Think about it this way. When you're working through a problem (especially one that feels a little uncomfortable or unfamiliar) the last thing you need is friction. You're already stretching. You're already curious or vulnerable or just trying to figure something out.

The moment you start translating your natural thoughts into some cold, stripped-down machine language, you've added a layer of work that has nothing to do with the problem you're trying to solve.

So here's a better mental model: talk to AI the way you'd talk to a competent new assistant on their first day.

You wouldn't be rude. But you also wouldn't over-perform. You'd say things like:

• "Hey, can you help me draft something?"
• "Thanks, let's pick this up tomorrow."
• "That's not quite what I meant. Let me try again."

Normal. Personal. No translation required.

Will "please" and "thank you" change what AI produces? Probably not much. But will forcing yourself to drop them make the whole experience feel clinical and weird? For a lot of people, yes. And weird is the enemy of useful.

Stay in your creative zone. Don't let the tool change how you think. That's backwards.

Why Shared Vocabulary Comes Before Better Results

Before we go any further, there's something worth naming.

People can't think clearly about what they can't name.

Vocabulary isn't academic. It's structural. We don't learn new words to sound smart. We learn them because they compress complex ideas, reduce confusion, and let us build on what we already know instead of starting from scratch every time.

That's especially true with AI.

Without a few shared terms, we either overcomplicate things, or avoid talking about them altogether. Neither one helps.

So, what follows is not a dictionary. It is not a glossary you need to memorize. It's a small, practical set of terms that will make your AI conversations clearer and calmer. Think of it as the shared language between you and the tool; enough to stay oriented, not so much that it becomes homework.

A Practical Vocabulary for AI Conversations

How We Talk to AI

Natural Language — This just means speaking or writing the way you normally would. Full sentences. Plain English. No special syntax required. AI is designed to work this way. This is what we just talked about above.

Context — Background information that helps AI understand what you actually need. Who you are. What you're trying to do. What constraints matter. The more relevant context you give, the better the output tends to be. Think of it as the difference between walking into a hardware store and saying "I need a thing" versus "I'm fixing a leaky faucet in my kitchen and I need a compression fitting." Same store. Completely different experience.

Iteration — Working in drafts. First pass, second pass, refinement. This is how AI works best — and honestly, it's how most good work gets done anyway. Expecting perfection on the first try is like skipping the blueprint and hoping the house stands.

How AI Produces Output

Generative AI — Systems that generate new content — text, images, summaries — based on patterns they've learned. When someone says "generative AI," this is what they mean. ChatGPT is one example. There are others.

Inference — The process of making predictions based on learned patterns. AI isn't recalling facts the way you and I do. It's inferring what's most likely to come next. There's a subtle but important difference.

Algorithm — A set of rules or processes the system follows to produce output. You don't need to understand the math. You just need to know that there is a process — it's not random, and it's not magic.

How AI Behaves in Practice

Hallucinations — When AI confidently produces information that sounds right but isn't actually true. Not lying. Not malicious. Just a reminder that judgment stays with you. Always verify anything that matters.

Agents / Agentic — AI systems that can take multiple steps or actions toward a goal on their own. More powerful. Also something to approach thoughtfully — like handing someone the keys before you've seen them drive.

Multimodal — Systems that can work across different types of input — text, images, audio. Useful. Still a tool.

The Horizon Term

AGI (Artificial General Intelligence) — A hypothetical future where AI can perform a wide range of tasks at a human level or beyond. Worth understanding conceptually. Not something you need to lose sleep over today. We'll talk about why in Part 3.

For more AI terms and definitions, visit our Vocabulary Systems page.

Why Natural Language Still Works (Even Though AI Isn't Human)

There's a tension worth addressing directly.

On one hand, AI is not a person. It does not understand meaning or intent the way we do. On the other hand, the best way to interact with it is still through natural conversation.

That is not a contradiction. It just requires a small shift in how you think about it.

AI does not understand you, but it's very good at modeling you.

When you speak naturally, you are not pretending AI is human. You're giving it better signal. What matters to you. What tone fits. What tradeoffs you care about. 

Here's the line that resolves the tension:

We speak to AI naturally not because it is human — but because we are.

 The more natural and contextual your input, the better the output.

You're Still the One in Charge

One last point. And it matters more than most people realize.

AI can draft. It can suggest. It can organize and summarize. But it cannot take responsibility.

You decide:

• What's true
• What's ethical
• What's appropriate
• What actually gets used

Shared vocabulary and natural language don't hand control to AI. They help you keep it.

That's the whole point.

What's Next

At this point, you understand what AI actually is and how to talk to it clearly.

And yet, for a lot of people, something still feels off. Even with understanding. Even with frameworks.

There's still unease. A sense that the ground is shifting faster than feels comfortable.

That's not a failure of learning. It's a human response to speed.

In Part 3, we'll talk about why this moment feels so overwhelming, why freezing or shutting down are completely understandable reactions, and why staying oriented matters more than keeping up.

No hype. No shame. Just perspective.

Part 3: Why AI Feels Overwhelming (And Why That Doesn't Mean You're Broken)

TLDR

• You don't need special syntax or perfect prompts. AI works best when you talk like yourself: clear, natural, and conversational.
• Should you say "please" and "thank you" to AI? The real question isn't whether it helps the AI, it's whether forcing yourself to talk like a robot helps you. Talk to AI like a competent new assistant on their first day.
• A small shared vocabulary (context, iteration, hallucinations, inference) helps you stay oriented without turning the interaction into homework.
• Natural language doesn't hand control to AI, it helps you keep it. You're still the one deciding what's true, ethical, and worth using.
• This is Part 2 of 4. Part 3 addresses why this moment feels so overwhelming, and why that doesn't mean you're broken.