AI training — Hamburg and the surrounding region
AI training in Hamburg — AI literacy instead of buzzword bingo.
In short
AI training under Article 4 of the EU AI Act teaches your staff what the tools in use actually do, which data may go in and how to judge a result. The foundation session runs half a day; with practice on your own tasks, a full day. For departments it is usually split into two half-days two weeks apart. Sessions are held separately by audience: users, managers, accountable roles – up to about twelve people where there is hands-on work. You receive attendance documentation with date, participant group, content covered and learning objectives, plus confirmations of attendance; the regulation does not ask for a certificate. As a rule the session takes place at your offices, at no surcharge in Hamburg and the surrounding region. Online works too. Sessions and materials are available in English.
Training after which your people work differently the next day. Not forty slides about neural networks, but real tasks from your own operation, solved on your own machines – and a clear statement of where the limits are.
The legal background
Why this landed on your desk in February 2025.
Article 4 of the European AI Act obliges every company that operates or provides AI systems to take measures supporting AI literacy among its staff. The obligation has applied since 2 February 2025, has no lower threshold for company size, and bites as soon as your people use ChatGPT or Microsoft Copilot for daily work.
The Digital Omnibus Regulation (EU) 2026/1744, in force since 27 July 2026, softened the wording: a particular level of competence no longer has to be guaranteed for each individual. What remains is the duty to do something at all – and to make it fit the systems, roles and risks actually present in the business.
What is required is therefore not a certificate but a measure that matches the actual deployment: staff should understand what the system does, what risks come with it and how to judge its output. That is exactly what this training is built around.
Content
What happens in the session.
-
Orientation without fog
What a language model actually does and what it does not. Why it sounds convincing when it is wrong. Which tasks become easier as a result – and which do not.
-
The tool landscape, sorted
Copilot, ChatGPT, Gemini, Claude, Mistral: what actually differs, and which criteria decide in a company. Legal area of processing, contract, integration, operating model – model quality comes last, not first.
-
Hands-on with your own tasks
Participants bring real work: a summary, a check, a draft, an evaluation. This is the longest part of the day and the only one that changes anything afterwards.
-
Data, secrets and the rules of the house
What may go in and what may not. Personal data, business secrets, non-disclosure obligations. A traffic-light sheet for the desk, handed out rather than read aloud.
-
Judging results
How to spot an invented source. Why arithmetic belongs in a calculator. What to check before something leaves the building – and how to keep checking once trust sets in.
Formats
Three audiences, three sessions.
One session for everyone is right for nobody. The split that has proved itself is simple.
Users
Everyone who works with the tools daily. What the tool can do and cannot, which data may go in, how to check a result, when to stop and ask a human.
- Half a day, or a full day with practice
- Up to twelve participants
- On your own systems
Managers
Everyone accountable for processes in which the tools are used. Where automation makes sense and where it does not, how to sample results, how to handle resistance.
- Half a day
- Small group, plenty of discussion
- Practical questions, not theory
Accountable roles
Management, data protection, works council, IT. Roles under the AI Act, high-risk classification, documentation duties, contracts, supervision.
- Two to three hours
- Shorter than the others – and indispensable
- Ends with a to-do list
The part nobody puts on the agenda
Shadow AI belongs in the room.
In almost every company, people are already working with private accounts. Not out of defiance – the work has to get done and nobody provided anything. A session that pretends otherwise loses the room in the first fifteen minutes. One that names it openly wins the room.
You do not have to approve of private use to take note of it. And the question „who here has already tried it?“ raises more hands in every room than the management expected.
Frequent questions
About the training.
How long does an AI training session take?
How many participants make sense?
Does the training take place at our offices?
Do we receive evidence for the EU AI Act requirement?
Our people are at very different levels. Is that a problem?
Can the training be held in English?
First conversation
A conversation that is worth something on its own.
Thirty minutes in which we go through your situation and I tell you what I would tackle first in your position — even if no engagement comes of it.