AI 101: a plain-language guide for people who never asked for AI

This series exists because of a real conversation. One of us was demoing Ethyx to a family member, and she had to stop the demo and say: "I really don't understand what AI is. You're speaking a different language."

She was right. Most explanations of AI assume you already care about it, already use it, or already work in tech. The good beginner resources that do exist — like the excellent free Elements of AI course from the University of Helsinki — are full courses that take 20 or more hours. That is a big ask when all you wanted was to understand what the app in front of you does.

AI 101 is our answer: a handful of short, honest reads. No jargon without an explanation. No homework. No hype — including about our own product. Most parts take about five to ten minutes; a few go longer when the topic needs it — each page shows an estimated reading time.

New here? Read Part 1 first. In a hurry? Jump to the cheat sheet below and come back when you have time.

Who this is for

• You have heard about ChatGPT or "AI chatbots" but never really used one.
• Someone in your life keeps showing you AI things and you nod politely.
• You have tried AI once or twice, got a weird answer, and quietly gave up.
• You use AI daily but wish you could explain it to your parents.

If you build AI systems for a living, this series is not for you — but you might know someone it is for.

The parts

1. What is AI, actually? — what you are really talking to when you talk to an AI, why it sounds so human, and the one rule that keeps you safe.
2. What should you actually use AI for? — three ways to think about what AI is for, with real examples from our house: allergy-safe recipes, 2am worries, and things that used to need a professional or a whole weekend.
3. How to ask so you get good answers — you do not need magic words, but a few habits make a big difference: say what "good" looks like, hand over the real thing, and push back.
4. When AI makes things up — why a confident answer is not the same as a correct one, the five places AI goes wrong most, why you can get a different answer each time you ask, and when that means AI is the wrong tool.
5. What happens to what you type? — where your words (and photos) actually go, who can see them, what to keep to yourself, and who owns what the AI writes back.
6. So many models — does it matter which one? — why there are different AIs with strange names, how to decode them, and why for everyday use the default pick is genuinely fine.
7. How tools make AI more useful — why web search, calculators, and other extensions fix the exact weaknesses covered in Part 4, and what they cannot fix.

All seven parts are published. You can also subscribe to the blog feed for future posts.

Cheat sheet

Print this page (use the Print cheat sheet button above) or bookmark it. These habits work in any chat app — not just Ethyx.

The one rule (Part 1)

• Confident tone is not proof. The AI sounds equally sure when it is wrong. Treat every answer like advice from a sharp stranger at a party — useful, not gospel.

Five briefing habits (Part 3)

• Say what "good" looks like (length, tone, audience, format).
• Hand over the real thing (paste the draft, email, or list — do not make the AI guess).
• Fill in context the AI cannot know (who this is for, what already happened).
• Push back when the first answer is off ("shorter," "warmer," "that fact is wrong — try again").
• Vague in, average out — every gap you leave gets filled with the most typical guess.

Four verification habits (Part 4)

• Ask twice on anything that matters (same question, fresh thread if needed).
• Open one source yourself on load-bearing facts (names, dates, prices, dosages).
• Use AI to understand; use official sources to decide.
• When the crowd is thin (niche facts, recent events, precise numbers), trust your skepticism more.

Privacy tune-up (Part 5)

• Use the nickname trick: "my friend" not her name, "Company A" not your employer, rounded amounts not account numbers.
• Keep work-confidential, medical, and financial identifiers out of personal chat apps.
• Assume your words are stored somewhere — the question is who can see them and whether they are used to train models. (See Part 5 for the full picture.)

Read this one first

• What is AI, actually? — start here if you have never used AI or still feel lost.

The promise

Everything in this series is the stuff we collectively wish someone had shown us at the start: the honest version, limitations first. Where AI is genuinely useful, we will say so. Where it fails, we will show you exactly how — because knowing how it fails is what makes it safe to use.

---

Ethyx is in closed testing with an access code today. You do not need Ethyx — or any particular product — for this series to be useful.