How to Become AI Literate in 30 Days: A Free Step-by-Step Roadmap

You can reach a working level of AI literacy in about 30 days at roughly 45 minutes a day. The sequence that works is: build the mental model first, then practise task framing, then train verification, then develop the judgement to decide when not to use AI at all. Skipping straight to prompts — which is what most courses do — is why most AI training does not stick.

This roadmap is structured around the four domains of the OECD/EC AILit framework — Engage, Create, Manage, Shape — one per week. It assumes no coding background. It works whether you are a teacher preparing for CBSE’s CT & AI curriculum, a student, or a professional whose employer has just handed you a licence and no instructions.

Before you start: the honest baseline

Write down your answers to these three questions today. You will re-answer them on day 30, and the difference is the actual measure of progress — not how many tools you have tried.

1. In two sentences, what is a language model actually doing when it answers you?
2. Name one task in your work that you should not give to an AI, and say why.
3. What happens to the text you paste into a chatbot?

Keep the answers. Do not look them up.

Week 1 — Engage: build the mental model

Goal: understand what AI is doing well enough to predict when it will fail. This is the week that makes the other three worth doing.

Day	What to do	Time
1	Learn one idea properly: a language model predicts the most plausible next piece of text, based on patterns in its training data. It is not retrieving facts from a database. Write this in your own words.	30 min
2	Make it fail deliberately. Ask an AI for five sources on a narrow topic in your field. Check every one. Record which existed, which were real but wrong in detail, and which were invented.	45 min
3	Learn the difference between rule-based systems and systems that learn from examples. Find one example of each in your daily life.	40 min
4	Understand training data. Ask an image generator for four occupations and observe the pattern in what comes back. Ask what data would produce that.	40 min
5	Learn the vocabulary you actually need — model, training data, hallucination, token, context window, fine-tuning, inference. Ten terms, not a hundred.	45 min
6	Compare three different AI tools on the identical question. Note where they agree, where they diverge, and what that tells you about confidence.	45 min
7	Checkpoint: explain to someone with no technical background why AI invents citations. If they understand it, you have week 1.	20 min

If you want the vocabulary in one place rather than looking terms up piecemeal, our AI glossary covers the terms in plain English.

Week 2 — Create: learn to frame tasks, not write prompts

Goal: get consistently useful output by learning to specify work clearly. The underlying skill is briefing, which is why people who are good at delegating are usually good at this immediately.

Day	What to do	Time
8	Learn the five elements of a good brief: role, audience, constraints, format, success criteria. Rewrite your worst recent AI request using all five.	45 min
9	Practise giving examples. Ask for something once with no example, once with one example of what “good” looks like. Compare quality.	40 min
10	Practise iteration. Take one output through three rounds of specific correction rather than restarting. Notice that specific beats polite.	45 min
11	Pick a real task you do weekly. Build a reusable brief for it. Save it somewhere you will find it again.	50 min
12	Use AI as a critic rather than a producer: give it your own work and ask for the three weakest points. This is usually more valuable than asking it to write.	40 min
13	Try one task AI is genuinely bad at — current local facts, precise arithmetic on a long list, anything requiring knowledge of your specific context. Notice the shape of the failure.	40 min
14	Checkpoint: produce one piece of real work with AI that you would be comfortable putting your name to, and note how much of it you changed.	60 min

Week 3 — Manage: verification, privacy and error-catching

Goal: stop being fooled. This is the week that protects you professionally, and it is the one most people skip. Microsoft’s 2026 research found that half of AI users now rate quality control of AI output as an increasingly important skill — it has become the differentiator.

Day	What to do	Time
15	Learn verification triage. Always check: numbers, names, dates, citations, legal or medical claims, and anything you are accountable for. Write this list down.	35 min
16	Take an AI-generated summary of a topic you know well. Mark every claim as verified, unverified or wrong. Count the ratio.	50 min
17	Practise tracing to a primary source. Take three AI claims and find the actual document behind each — not a blog citing a blog.	50 min
18	Data hygiene audit. List what you have pasted into an AI tool in the last month. Flag anything containing a person’s name, a phone number, a student record or client data.	40 min
19	Read your organisation’s or school’s AI policy. If there is none, draft five lines: what is allowed, what is banned, what must be disclosed, what data never goes in, who decides exceptions.	45 min
20	Learn where AI confidence and AI accuracy come apart. Ask a question you know the answer to, in a field you know well, and note how confident a wrong answer sounds.	35 min
21	Checkpoint: have someone give you an AI-written document with three deliberate errors planted. Find them.	40 min

Week 4 — Shape: judgement, ethics and redesign

Goal: move from using AI to deciding how it should be used — by you, and by the people around you.

Day	What to do	Time
22	Write your personal “do not automate” list — the three tasks where the struggle is the point, or where you could not evaluate the output.	40 min
23	Study one real case of AI bias in hiring, lending or policing. Identify what in the data caused it.	45 min
24	Work out your disclosure rule: when would a reader’s judgement of your work change if they knew AI was involved? Disclose there.	35 min
25	Map one of your workflows end to end. Mark each step: keep human, AI-assisted, or fully delegated. Justify each choice in a sentence.	50 min
26	Teach it. Explain one AI concept to a colleague, a student or your child. Teaching exposes the gaps nothing else does.	40 min
27	Learn the governance basics relevant to you — for schools, CBSE’s approach to AI and bias from Class 7; for companies with EU clients, Article 4 of the EU AI Act, in force since 2 February 2025.	45 min
28–29	Do one substantial real project using everything: framed brief, AI draft, verified claims, documented decisions about what stayed human.	2 × 60 min
30	Final checkpoint: re-answer the three baseline questions. Compare with day 1.	30 min

Where you should be on day 30

A realistic outcome, stated honestly. After 30 days you should be solidly at the Engage and Create levels and functional at Manage. You will not be at Shape — that level depends on judgement built through repeated real decisions over quarters, not weeks, and anyone promising it in a month is selling something.

Concretely, you should be able to:

• Explain why AI hallucinates, without jargon
• Write a brief that reliably gets usable output first time
• Spot which claims in an AI answer need checking, and check them
• State what you will never paste into a chatbot
• Name three tasks you deliberately keep human, and defend the choice

Those map onto five of the eight core AI literacy skills. The remaining three — bias recognition, attribution norms and workflow redesign — are the work of the following quarter.

Adapting this for a school or a team

Two adjustments make this work at scale rather than individually.

• Stretch it to a term, not a month. One exercise a week for 12 weeks beats a daily commitment nobody keeps in an exam term.
• Have managers or heads of department do it visibly. Microsoft’s 2026 data found organisational factors accounted for 67% of AI impact against 32% for individual behaviour, and that reported value rose 17 points when managers actively modelled AI use. A programme leaders do not visibly participate in produces compliance, not literacy.

For CBSE schools, the natural place to anchor this is the CT & AI curriculum itself — teachers building their own literacy on the same four-domain structure they will be teaching from Classes 3 to 8.

Frequently asked questions

Can I really become AI literate in 30 days?

To a working level, yes — roughly 20 to 25 hours of deliberate practice gets you to the point of using AI productively and catching its errors. Full literacy including bias judgement and workflow redesign takes longer. The 30 days gets you past the dangerous stage, which is the point.

Do I need to pay for AI tools to do this?

No. Every exercise here works on free tiers. A paid subscription changes what you can produce, not what you can learn about how these systems behave.

I have no technical background. Is this still for me?

Yes, and non-technical learners often do better on weeks 3 and 4. Verification and judgement are closer to editing and professional scepticism than to engineering. Technical users tend to over-trust systems they feel they understand.

What if I miss days?

Continue rather than restart. The sequence matters far more than the calendar — week 3 is only useful once week 1 is genuinely absorbed. Taking eight weeks over it changes nothing.

Is there a certificate at the end?

Not for this roadmap — it is free and self-directed. If you need a credential for an employer or a school file, structured programmes with assessment exist, including PiyushAI’s AI training for teachers and staff. Be sceptical of any certificate awarded without a performance-based assessment.

How is this different from a prompt engineering course?

Prompt engineering occupies week 2 of this roadmap. The other three weeks — understanding, verification and judgement — are what determine whether AI helps you or embarrasses you, and they are what prompt courses generally omit.

---

PiyushAI runs structured AI literacy programmes for teachers and school staff, for schools implementing CBSE’s CT & AI curriculum for Classes 3–8, and for working professionals. Founded by Piyush Wairale (IIT Madras).

Sources: OECD/EC AILit Framework; Microsoft 2026 Work Trend Index; CBSE CT & AI Curriculum 2026–27; EU AI Act, Article 4.