Understand AI
How generative AI works, where its limits lie and why plausible answers can still be wrong. Recognise hallucinations and bias through clear examples.
We make AI understandable and useful in everyday work. Our training fits your tools, tasks and people, from initial use to clear rules for responsible application.
Book an appointmentAI literacy goes beyond good prompts. Your team should recognise opportunities, understand limitations and evaluate outputs professionally. We connect the foundations with actual tasks from your business. Whether drafting, translation, knowledge retrieval or initial analysis, we practise where AI can help and where a person must check or decide.
These modules offer a possible training plan. We agree on selection and depth according to the tools actually used and your team’s needs.
How generative AI works, where its limits lie and why plausible answers can still be wrong. Recognise hallucinations and bias through clear examples.
Define tasks clearly, provide context and quality criteria, and improve outputs. Practise suitable tasks such as drafting, translation and structured summaries.
Consider confidential information, personal data and copyright questions during use. Which data can enter which tool, and when is approval needed?
Check claims against suitable sources, recognise professional errors and define review steps. AI supports work; responsible people retain ownership.
Record permitted tools, responsibilities, boundaries and escalation routes in clear language, so your team can act consistently when uncertain.
Review exercises, gather questions and align refreshers with actual use. New tools or changed tasks provide reasons to update content.
Article 4 of the EU AI Act also applies in Germany. As amended by the AI Omnibus, providers and deployers must take measures to support the development of AI literacy among people working with AI on their behalf. They do not have to guarantee a particular level of competence for any individual. Measures take account of tasks, prior knowledge, the systems used and their risks.
| Question | Guidance |
|---|---|
| Who is covered? | Staff and external personnel operating or using AI on your business’s behalf. |
| Does ChatGPT count? | Yes. This includes uses such as advertising copy and translation. |
| One course for everyone? | No. Content and depth should suit the use case and audience. |
| A certificate or fixed training hours? | Article 4 requires neither a certified standard course nor a fixed number of training hours. |
| Internal or external measures? | Both are possible: briefings, workshops, self-directed learning and guided training. |
| Documentation? | The Bundesnetzagentur recommends recording measures clearly and refreshing them according to need. |
Which AI tools do you already use? Who uses them, for which tasks and with what data? We identify existing knowledge, uncertainties and goals.
We define learning goals, content, exercises and a suitable format. Introductory, practical and role-specific sessions can be combined as needed; scope and dates are agreed individually.
Your team tries the tools using agreed examples. We discuss outputs and mistakes together. Exercises can use anonymised or purpose-created data.
Content, date, scope and attendance are recorded clearly. Together we identify open questions and useful next learning steps or refreshers.
Your team receives accessible foundations and concrete guidance for its own AI use.
Suitable exercises, review steps and agreed usage rules support transfer to everyday work.
An attendance record can document content and scope. It is not official certification or a blanket guarantee that all business obligations have been fulfilled.
No. The relevant people are those working with AI on your business’s behalf. Measures should suit their role, prior knowledge and the risks. Management, marketing, service and specialist departments may need different levels of depth.
Article 4 does not prescribe special trainer authorisation or an IHK, TÜV or INQA qualification. A suitable training concept matters. Selection should reflect professional suitability, the systems used and learning needs.
That depends on use. AI tools, tasks and risks change. We plan with you how to address practical questions and refresh content when needed. Article 4 does not prescribe a fixed annual number of hours.
Yes, training content and scope can be documented in an attendance record. The business remains responsible for appropriate measures. A record is not an official seal of approval and does not replace assessment of the particular use.
Training develops knowledge and practical skills. INQA coaching supports a participatory organisational change process and has its own funding conditions. AI training is not INQA-funded merely because of its topic.
More about INQA coachingBook a conversation. We discuss tools, tasks, participants and goals, then develop a suitable concept and offer without presenting a standard package as legally mandatory.
Book an appointmentChecked on 22 October 2026. The legal information provides orientation; training scope and measures depend on your particular use.
Bundesnetzagentur: AI literacyEuropean Commission: AI literacy questionsLet us explore what your team already knows and which learning steps your business needs next.
Book an appointment