Quick answer: AI education in 2026 sits in an uncomfortable middle: most universities and schools have rolled out something, but content quality and assessment discipline vary widely. Enterprise training has matured fastest because the EU AI Act Article 4 gave it a compliance hook.

What changed in 2025–2026

The EU AI Act literacy obligation came into force in February 2025. That single legal anchor redirected enterprise L&D budgets into structured AI training for the first time. Schools followed, mostly through government guidance rather than curriculum reform.

Where it works

Short, role-specific training for regulated staff (HR, clinicians, educators, developers).
Gamified micro-learning with assessment — AI Quest, Elements of AI, Code.org AI.
Vendor-agnostic content. Tool-specific training dates fast.

Where it fails

Multi-hour videos with no assessment — the old L&D default.
Generic “what is AI” courses aimed at cabinets, not end-users.
Content written in 2023 and never updated; model behaviour has changed materially since.

Where the market is growing

Enterprise compliance packages (£30–£100 per seat per year).
K-12 classroom curricula funded by national education budgets.
University modules embedded in non-CS programmes (business, medicine, law).

What schools are adopting

No single standard has emerged. UK schools reference the DfE framework; Singapore schools reference MOE Digital Literacy; EU schools reference DigComp 2.2. US adoption remains district-by-district.

What the next year looks like

Expect national certifications, more integration into professional registrations (medicine, legal, teaching), and further pressure on content freshness. The winners in this market will be the providers who update curricula monthly rather than annually.

State of AI Education 2026: Where Teaching AI Actually Stands