EU AI Act Readiness Checklist
A practical checklist for engineering and product teams whose AI systems reach people in the EU. It walks through whether the AI Act applies to you and in which role, the prohibited practices, high-risk classification, transparency duties, general-purpose AI model obligations, deployer and provider duties for high-risk systems, and the dates each part applies, with a severity and a verification step for every check.
7
Sefism · Oct 2026
32
Sefism · Oct 2026
8
Sefism · Oct 2026
AI
Sefism · Oct 2026
Checklist
AI & Machine Learning
Sefism
20 min read
Intermediate
Free
What this checklist is
A practical checklist for engineering and product teams whose AI systems reach people in the EU. It walks through whether the AI Act applies to you and in which role, the prohibited practices, high-risk classification, transparency duties, general-purpose AI model obligations, deployer and provider duties for high-risk systems, and the dates each part applies, with a severity and a verification step for every check.
It has 7 chapters and 32 checks, each with how to verify it and how much it matters. For teams building or deploying AI that reaches EU users.
Chapter by chapter
1. Scope and your role (5 checks)
Before any obligation, work out whether the Act reaches your systems and in what role. Providers and deployers carry very different duties, and one company can be both.
2. Prohibited practices (4 checks)
Article 5 bans a short list of practices outright. These apply already, and carry the highest fines, so screen every use case against them first.
3. High-risk classification (4 checks)
Whether a system is high-risk decides most of your workload. Classify each system against Article 6 and Annex III, and record the reasoning.
4. Transparency duties (4 checks)
Article 50 covers chatbots, synthetic content, deepfakes and some biometric uses, whatever their risk level. According to the Commission, these rules come into effect in August 2026.
5. General-purpose AI models (4 checks)
Most app teams use general-purpose models rather than provide them, but both sides have work to do. These checks cover Article 53 duties and what to ask your model vendor for.
6. High-risk systems: provider and deployer duties (6 checks)
If a system is high-risk, these duties apply from 2 December 2027 for Annex III use cases and 2 August 2028 for Annex I products. They take time to build, so plan them now.
7. AI literacy, governance and dates (5 checks)
The last checks keep the work going: staff literacy, ownership, a dated plan and a way to notice when the rules change.
How it works on Sefism
The checklist is one page: each check has a severity (must, should or nice to have) and a concrete way to verify it, and members can tick checks off as they go. The opening chapters are open to everyone. Adding it to your account uses credits, opens every chapter and keeps your progress.
Every claim is checked against the sources listed on the page (8 cited).
Ready to start?
Opens on Sefism in a new tab.
Stuck on something specific?
Writing only gets you so far. If you want an answer to your situation rather than the general case, book a session and we will work through it together. Sessions are free for approved Sefism members, and a few slots open each week.
Follow along
New writing, resources and project ideas land here first.