Sample — excerpt from the full Kramer Consulting seminar · kramerconsulting.lu/training
DLH Belval AI Training Programme · Module ACT-A

AI Act
in 4 Hours

No legal background needed. Leave knowing where your organisation stands.

22 March 2026  ·  Beginner  ·  4 Hours  ·  Max 20 participants

⚠ Partially in force since 2 August 2025
Prohibited practices are already banned. AI literacy obligations have started. Enforcement of high-risk systems begins August 2026.
Today's Programme

What we cover in 4 hours

  1. 01 The Risk Pyramid — four tiers, your obligations
  2. 02 Prohibited Practices — what is already banned
  3. 03 Deployer Obligations — what your organisation must do
  4. 04 Your Next Steps — actionable priorities for this week

Three hands-on exercises · One compliance checklist · No jargon

By the end of this session, you will be able to

DateWhat it means
2 Aug 2025 ✓Prohibited practices banned. AI literacy obligation begins.
2 Aug 2026High-risk AI obligations fully apply. Fines of up to 7% of global annual turnover.
2 Aug 2027AI systems already on the market must be brought into compliance.
Part 01 of 03

The Risk
Pyramid

The EU AI Act classifies every AI system into one of four risk tiers. Your compliance obligations depend entirely on which tier applies to the systems your organisation uses.

The Four Risk Tiers

Your obligations depend on which tier your AI systems fall under
Unacceptable Risk
🚫 Banned outright. Do not use.
Prohibited under the AI Act with no exceptions for commercial use. Social scoring, emotion recognition in workplaces, real-time biometric surveillance.
High Risk
⚠ Heavily regulated. Verify before you deploy.
Permitted but requires conformity assessment, human oversight, and fundamental rights impact assessment. CV screening, educational assessment AI, benefits eligibility.
Limited Risk
ℹ Transparency obligation only.
You must inform users they are interacting with an AI system. No other specific obligations unless the system also falls under a higher-risk category. Chatbots, AI-generated content, virtual assistants.
Minimal Risk
✔ No obligations. Free to use.
The vast majority of AI tools fall here. Spam filters, auto-complete, recommendation engines, grammar checkers, most productivity tools.
Exercise 1 · Try it now · 20 min

Traffic Light Sort

For each scenario, assign a risk tier and explain your reasoning to a partner
Scenario A
Your HR department uses an AI tool that automatically ranks job applications and filters out candidates before a human sees them.
Scenario B
Your website has an AI chatbot that responds to customer queries. It looks and responds like a human.
Scenario C
A government agency rates citizens on a trustworthiness score based on their online behaviour.
Scenario D
Your organisation tracks employee keystrokes and generates a weekly productivity score shared with managers.
Answers: A → High Risk (Domain 4 employment) · B → Limited Risk (transparency) · C → Unacceptable (social scoring) · D → High Risk (worker monitoring)
Sample

This is a sample.

The full seminar is hands-on and runs at the Digital Learning Hub, Belval, or in-company.

Seminar details & dates →