The EU AI Act, for people who build the product.
A plain-language, designer-first read on Regulation (EU) 2024/1689: how the risk tiers work and the interface patterns behind the common obligations, so you can design for them from the start.
Free for one project. Educational only, not legal advice: this guide does not classify or certify any product.
- Prohibited
- High-risk
- Limited risk
- Minimal risk
Many of the Act’s obligations show up in the interface.
Human oversight, disclosure and record-keeping are, in practice, design problems. They are easiest to get right when the team asks about them in discovery.
Found at launch
Each pattern becomes a patch.
- An override that arrives after the action ran
- A disclosure buried in the footer
- Logs that no one can read or export
Designed in from discovery
The same three patterns, planned early.
- A person can step in while the decision still matters
- People know when they are dealing with AI, at the moment of use
- A review surface a team can actually use
Four tiers, from banned to barely touched.
The Act scales its demands to the risk a system poses. Here is what each tier means in general. The examples are illustrative, not a ruling on any product.
Prohibited
A short list of practices the Act bans outright, set out in Article 5.
For example: Social scoring that ranks people by their behaviour and limits their access to services.
High-risk
Allowed, with the Act's most demanding duties before and after going to market, such as risk management, documentation, human oversight and logging.
For example: An AI system that screens job applicants or ranks candidates.
Limited risk
Allowed, with disclosure duties in specific situations, so people know when they are dealing with AI or AI-generated content.
For example: A customer-service chatbot, which generally has to make clear it is an AI.
Minimal risk
Few or no specific duties under the Act. Other rules can still apply, and voluntary codes of conduct are encouraged.
For example: An AI-enabled spam filter.
Wondering where your system sits? Do not take a verdict from a web page. Work through an established self-assessment and confirm anything that matters with counsel.
The interface patterns behind the obligations.
These are common patterns teams reach for. They are design guidance, not a guarantee of compliance.
Human oversight & intervention
Supports Article 14 (human oversight)
Give a person a real, timely way to understand, question, and override what the AI is doing — before the consequence lands.
A duty for high-risk systems.
Watch out: An override that only arrives after the action executed isn't oversight. Put the decision in front of the human while it still matters.
Transparency & disclosure
Supports Articles 50 & 13 (transparency, instructions)
Make it unmistakable when someone is dealing with AI — and honest about what it can and can't do.
Disclosure duties target specific systems, such as chatbots and AI-generated content; instructions for use apply to high-risk systems.
Watch out: A disclosure buried in the footer, or shown once at onboarding, doesn't land. Keep it present at the moment of the interaction.
Logging & auditability
Supports Article 12 (record-keeping)
Design the surfaces that let a team see what the system did — and prove it later.
A duty for high-risk systems.
Watch out: Logs no one can read or export aren't evidence. Design the review surface, not just the database table.
Plan the patterns in a project, not a PDF.
Sign up to plan how each pattern shows up in your product, track its status, attach evidence, and work through a discovery checklist with your team.
Design-pattern planner · sample
- In design
Human oversight & intervention
Override control and pause, shown with confidence.
- Planned
Transparency & disclosure
AI badge in the chat, limits shown in context.
- Not started
Logging & auditability
Add notes and evidence links: the PR, the design file, the policy.
Discovery checklist
7 questions, in 3 groups.
- Scope & role
- Risk & rights
- Design surfaces
Questions worth asking in discovery or sprint planning. A prompt for your team’s own thinking, not an assessment.
Prefer a file? Download it (Markdown) →Start free. Upgrade when you need more.
Free
- Every tool, in full (free plan limits apply)
- One project per tool
- One survey round, up to 25 responses
Pro
for your first 3 months, then €29/mo.
- Unlimited projects
- Unlimited survey rounds & responses
- Historical trends across rounds
- AI coach & assisted canvases across the framework
Read the text itself.
The Act phases in over several years and application dates can change, so this page does not list them. Check the official sources before relying on any date.
Article by article (unofficial site)
Educational information only. This guide explains how the EU AI Act is structured. It is not legal advice and does not classify or certify any specific product. For decisions that carry legal weight, consult qualified counsel.