I Rebuilt AI For Everyone

In 2017 I sat with the managers of small and medium companies who were eligible for our 100 Experiments programme and watched them hesitate. Their question was never whether AI worked. It was whether AI would replace them. We could not build a national AI programme on top of that fear, so we built the antidote first: a short course to demystify AI for absolutely everyone.

AI For Everyone launched in the second quarter of 2018 as a three-hour face-to-face workshop, typically drawing between one hundred and four hundred people a session. In December 2019 we put it online, and the polytechnics and the Institute of Technical Education adopted it as a foundation module. Through our literacy programmes, more than 300,000 Singaporeans have now been reached.

Last month I tore it down and rebuilt it.

What came down

The teaching was never the problem. The delivery had quietly collected the assumptions of 2019 online learning. The course ran on WordPress, the way most courses run: plugins, a page builder, PHP, a login screen, an analytics stack. Every learner had to create an account before they could learn what AI is. A literacy course that asks for your email before it demystifies AI has already taught one lesson, and it is the wrong one.

So the new AI For Everyone is HTML and TypeScript on Cloudflare. No plugins, no page builder, no PHP. Static pages, fast, and they work on a bad connection, because a course for everyone has to survive everyone's network. The train tunnel, the shared data plan, the branch office at the end of a long line: those are the networks a national literacy course actually lives on.

No signup, no tracking

There is no account. You never sign up. You arrive, you learn, you leave.

The course tracks nothing and calls nobody. Your progress stays in your browser, and you can carry it in the link: finish a lesson at the office, paste the link into your phone, continue on the bus. We cannot leak what we never collect, and after a decade of watching course platforms turn into data harvesters, I did not want our literacy layer to be another one.

A Try-it in every lesson

The bigger change is inside the lessons. Every lesson now ends with a hands-on Try-it, and the Try-it is done on your own real work, not on a toy dataset. You apply whatever the lesson taught, a few sentences, a spreadsheet, a real email you have to send, to the work in front of you.

This comes straight from everything we have learned building AI Singapore's talent programmes. Watching a video builds awareness. Doing the thing on your own work builds capability. The gap between those two is where most corporate AI training quietly fails, and the Try-it is my answer: small, immediate, yours.

Readiness, measured

The course now sits on the AIRI readiness thread end to end. Before you start, take the personal assessment, pAIRI, at airihub.io. It places you on the L0 (AI Unaware) to L2 (AI Ready) and to L4 (AI Catalyst) AI Readiness ladder across the five readiness dimensions. Bring your result into AI For Everyone and the course tracks your progress against your readiness, so you can see yourself move rather than just accumulate completion ticks. Organisations can do the same with oAIRI, the organisational assessment, and watch the readiness of their people shift as the course is taken.

I built the AI Readiness Index to make AI readiness measurable. A literacy course that cannot show movement against that measure is just content.

Built to travel

AI For Everyone and the AIRI Framework has always travelled. Egypt took it first, translating the AI4E course into standard Arabic for its citizens under a government-to-government agreement. Others through my work with the Global Partnership in AI (GPAI) now under OECD, adopted AIRI and a SME version of AI4E and it is now running in Serbia, Poland and France. The old WordPress bundle made each of those journeys heavier than it needed to be. A static course with no accounts is closer to a folder and a domain, which makes the next translation, and the next country, a smaller conversation.

The voice, and the Monkey King

All 35 narrations were re-recorded in my voice, punchier than before, and the welcome lesson leads the update. It keeps the Journey to the West spine it has always had, with new art and a new video. If you have read this newsletter before, you know the Monkey King already: the same Sun Wukong who retired my Avengers slide now opens the course, plucked hairs and all, because the story of one person multiplying their own capability is the most honest story in AI right now.

And yes, this was built the way I keep saying things can be built now: one person, plus AI agents, evenings and weekends, no grant and no team.

The lesson for organisations

If you are responsible for your organisation's or country's AI literacy, the expensive assumptions are worth naming. A literacy course does not need sign-ups and accounts, an LMS licence to be credible, or an analytics stack to understand the customer, or a course fee. AI literacy is a basic skill. It should be free for everyone to acquire. Privacy can be architecture rather than a policy paragraph. Static pages cost almost nothing to serve at national scale. If a course for 300,000 people can run this lean, your onboarding portal can too.

AI For Everyone is free, at aiready.sg. Take pAIRI at airihub.io first, then start at lesson one. The Monkey King is waiting.

By Laurence Liew — Director, AI Innovation, AI Singapore | Author of AI-First Nation | Co-Chair GPAI I&C WG.