AI driving licence: what is behind it, and do employees need one?
“AI driving licence” translates the German marketing term “KI-Führerschein”, used for basic AI courses. It is not a protected qualification title and not regulated by law. The EU AI Act has no such licence; Art. 4 requires measures on AI literacy.
What is an AI driving licence?
An AI driving licence is not an official qualification. In Germany, course providers sell basic AI courses under the name “KI-Führerschein”, mostly online and often with a confirmation of participation at the end. The name is not a protected qualification title, and content and scope are not uniformly regulated: one course can take two hours, another two days. Unlike a car driving licence, there is no official exam, and no law prescribes it.
Does the EU AI Act require an AI driving licence?
No. For AI literacy, the EU AI Act provides for neither a driving licence nor an exam or a certificate. Art. 4 requires something else: providers and deployers take measures to support the development of AI literacy of their staff and other persons acting on their behalf (Art. 4(1)). Which measures these are depends on the knowledge, experience, education and training of the people involved and the context of use; the provision prescribes no particular format.
In its questions and answers on AI literacy, the European Commission makes clear that Art. 4 does not require a certificate. An internal record of the training and other measures is enough.
What a basic AI course should cover
Whether a course is called “AI driving licence” says little. The content is what matters. A basic course for everyone who uses AI at work should cover these topics:
- Basics: what an AI system is, how it works in principle and why data quality matters.
- Risks: typical errors and biases in AI outputs, and the risk classes of the EU AI Act.
- Obligations: what Art. 4 requires, who is a provider and who a deployer, plus human oversight and transparency.
- Data protection: which data belongs in AI tools and which does not.
- Effective prompting: formulating tasks clearly and checking results critically.
The AI-Konform training covers these topics in four lessons, each with a short quiz, about 30 to 60 minutes per person, in 26 languages. More on the page EU AI Act training.
What companies should look out for
Anyone who wants to introduce or buy an AI driving licence for employees should check four points:
- Documentation per person: record who completed which measure and when. The personal certificate of participation for each person and the audit report from AI-Konform document participation and the knowledge test; they are not an official certificate.
- Languages: teams with several native languages need the content in the language of their employees.
- Effort: time per person and costs for the whole team belong in the calculation. The page cost of EU AI Act training shows market prices.
- Advertising claims: a course that promises a “legally required” or “official” AI driving licence promises more than the law provides.
Frequently asked questions about the AI driving licence
Do employees need an AI driving licence?
No. The EU AI Act does not prescribe an AI driving licence and, for AI literacy, requires neither an exam nor a certificate. Art. 4 requires measures on AI literacy; whether that is a basic course, an introduction to the systems in use or another format is for the company to decide. A documented training course is a simple measure, more on the page AI literacy obligation under Art. 4.
Who issues an AI driving licence?
The course and the certificate come from the respective course provider, such as a training company, a chamber of commerce or the employer itself. The term is not a protected qualification title, and no law regulates content, exam or recognition. Its value therefore depends mainly on what is taught and how the knowledge was tested.
Is an AI driving licence enough to implement Art. 4?
A basic course can be a suitable measure, but it is not automatic. Art. 4 requires measures that take into account the knowledge, experience, education and training of the people involved and the context of use. A general course does not necessarily cover the systems actually in use. It therefore makes sense to combine a basic course for everyone, an introduction to the systems in use and a record of who took part and when.
Train your team now What the training covers
Note: this article is for general information only and does not constitute legal advice.