We've open sourced the UAIL (Universal AI Level)
Sophie published the Universal AI Level today: a 0 to 10 scale for what someone can actually do with AI, not which tools they have opened. Eleven levels: 0 for never touched it, 10 for building the models yourself, and the no-code ceiling sits at level 5.
It exists because of the workshops: we needed a way to tell people which level a session was pitched at, so they could book the right one. "Beginner" and "advanced" mean something different in every room that uses them.
I am the technical half of Inkie, so my instinct was predictable: it went on GitHub that same evening, under a CC BY 4.0 licence.
The quick-reference card:
Two strangers, one number
A scale only works if two strangers read the same definition and land on the same number. That means the definitions have to be visible and quotable by people who have never heard of Inkie, with nothing resting on whether we keep a page online. Skip that and you have shoe sizes: a 9 in one shop is a 7 in the next, and everyone learns to ignore the number and try things on.
GitHub does the visibility part better than any page we could host. Pin a version, diff what changed, fork it, open an issue when a level stops matching the tools. All of that comes free the moment the spec lives in a repo.
Use it, credit it, don't rebadge it
The licence is CC BY 4.0. Use it, share it, adapt it, build a paid product on top of it, no permission needed. It asks two things back: credit it (the README has a suggested line), and do not present a modified version as the official UAIL.
What's actually in the repo
SPEC.md is the normative text: every level with its full description, a plain-English version, and a "how do you know?" test. If a PDF and the repo ever disagree, the spec wins.
data/levels.json is the same data as JSON. Nobody has built anything on it yet, but it is there for whoever wants a self-assessment quiz or a course picker without parsing a PDF first.
Most people are never going to open a repo, so five PDF guides ship alongside it: the 12-page overview, a quick reference, a course labelling guide, an employer guide, and a printable self-assessment.
Where the workshops fit
The workshops are why the scale exists, and they are starting now: sessions that train people in AI, each pitched at a stated level so you can book the right room. We can run them anywhere from level 0 to level 8. The sweet spot is the stretch from level 1, where a lot of people sit today, up to level 5, the top of the no-code range, because that is where most people are stuck. Moving people up the scale is the job.
The ceiling will move
It is numbered v1.0 because the thing being measured does not stay still. The no-code ceiling is drawn at level 5 against August 2026's tools; some of what needs code today will not need it next year, and the line will have to move when it does. That is the main reason it is open source at all. CONTRIBUTING.md and CHANGELOG.md are there for exactly that job: version the changes in public, instead of quietly editing a page and hoping nobody screenshotted the old one.
The repo is at github.com/simon-inkie/uail. The live version is at inkie.ink/universal-ai-level. If the wording is off somewhere, or you want to add something, a pull request gets it fixed faster than an email.
If you use UAIL to label a course or run a hiring round, let me know how it goes. And in the meantime, go out and make some cool shit.