From zero to AI master.
30 days. 5 minutes a day.

Yes. Zekai is an independent AI tools directory — the challenge is how we introduce people to the tools we test. Some tool pages contain affiliate links; the challenge itself costs nothing and requires no purchase.

No, and in the areas that matter most it legally cannot. The AI Act treats access to essential public services and benefits, law enforcement, migration and the administration of justice as high-risk, and requires that deployers ensure staff are trained to exercise human oversight. This challenge is about the work around the decision: correspondence people can understand, briefings that get read, consultation analysis that keeps the minority view, and casework preparation. The decision, and the duty to explain it, stays with a named official.

Installed on day 6, and the public sector is where the law draws its hardest lines. Since 2 February 2025 the AI Act has prohibited outright a set of practices that fall largely on public bodies: social scoring, predicting an individual risk of committing a crime based solely on profiling, untargeted scraping of facial images to build recognition databases, emotion recognition in work and education, certain biometric categorisation, and real-time remote biometric identification in public spaces for law enforcement. Above those sit the high-risk categories where most public sector work lives. The practical rule underneath all of it: a named official must be able to explain and justify a decision about a person.

The EU rules are used here because they are the most detailed and specific publicly available framework, and because they are already in force. They are not the only ones and they may not be yours. What transfers to any jurisdiction is the underlying reasoning: which uses are so risky they should not happen at all, which affect a person rights enough to demand human oversight and explanation, and what a public body must be able to show afterwards. Day 6 asks you to locate your own work against those categories, and the challenge repeatedly tells you to confirm what applies where you are. Nothing here is legal advice.

That is normal, and day 18 treats it as a governance signal rather than misconduct. Staff use their own devices when nothing has been provided or permitted, and responding with a prohibition produces concealment rather than compliance. The organisations handling this well establish honestly what people are already doing, are clear about what is genuinely prohibited and why, and provide a sanctioned route quickly. Day 20 gives you the do-not-paste list to publish in the meantime, and day 24 turns it into training built around failure modes rather than features.

Because binge-reading builds nothing. The challenge unlocks one drill per day from the day you sign — the way any skill gets built: daily reps. Day 1 is open right now so you can try it before committing.

Zero. Day 1 assumes you have never written a deliberate prompt. By day 14 you will know what your organisation must record when AI contributes to a decision.

Life happens, and so does an unplanned stop. The challenge is sequential, not strict-daily. Pick up where you left off; the certificate cares that you finish all 30, not that you were perfect.

This is the Government & Public Sector Edition. Doctors, lawyers, accountants, teachers, farmers, video editors and many more are already running theirs, with more coming. Sign up and we’ll tell you when yours goes live.