Using AI at university or work: what’s allowed in Germany and the EU

Students: the assignment rules matter most

For education, the European Commission has issued guidance on the ethical use of AI and data in teaching and learning . That matters because it shows AI in education is not treated only as a ban-or-allow question. It is also a question of ethics, transparency, data protection and responsible organisation.

For a student, though, the real question is rarely abstract. It is not whether AI is allowed somewhere in the EU. It is whether you may use this tool for this essay, exam, presentation, lab report or thesis. That answer comes from the local rulebook: the assessment regulation, module description, course page, university policy or direct instruction from the lecturer .

For universities and research institutions, the sources highlight transparency, data protection, security, internal guidelines and training as central requirements or implementation tasks . KI:edu.nrw, a German higher-education AI project, also frames the AI Act as a set of requirements that universities must take into account and implement step by step .

In practice: before submitting graded work, check whether AI use is allowed, restricted, must be disclosed, or is excluded. If the rule is unclear, ask before you submit. A short question in advance is safer than a long explanation afterwards.

Employees: a public AI tool is not automatically work-approved

At work, the same broad principle applies: the sources point to regulation, not a total ban. Germany applies the EU AI Act as its primary AI-specific framework, while other German or EU rules may also matter depending on the sector, the data and the context .

For employees, that means a free, public or widely used AI tool is not automatically approved for work. The key issues are whether the use has been cleared internally, what data you enter and whether data-protection or security rules are involved .

Be especially careful with personal data, customer data, internal documents, confidential information or security-sensitive material. The sources on universities, research and organisational AI use repeatedly point to data protection and security as core requirements .

Transparency can also become an issue in business settings. Examples raised in the source material include AI chatbots on a website, automated AI replies to customers and published content where it is not clear that AI generated it .

What the EU AI Act changes in practice

One practical change is AI literacy. From 2 February 2025, a central AI Act provision applies: covered providers and deployers of AI systems must ensure that appropriate AI competence is in place . This is not a blanket ban on individual users. It is a signal that organisations should not let AI use develop without rules, training or oversight.

The requirements become more serious when AI affects assessments or decisions. In higher education, one source gives the example of an AI system that supports the grading of student work; because it can directly influence assessment processes in education, it may fall into a high-risk category . That does not mean every writing, learning or research assistant is automatically high-risk. The purpose, context and effect of the system matter.

Quick check before using AI

Use these questions before relying on an AI tool for study or work:

1. Is there a written rule? At university, check assessment rules, course instructions or university guidance. At work, check IT, data-protection, compliance or AI policies .
2. Do you need to disclose the use? Look for any requirement to mark or explain AI assistance, especially for assessed work, customer-facing communication or published content .
3. What data will you enter? If the material involves privacy, confidentiality or security, a casual assumption that the tool is fine is not enough .
4. Could the AI affect a grade, assessment or decision? Systems that influence educational assessment can face stricter treatment; AI support in grading student work is specifically described as a possible high-risk example .
5. Has the organisation approved and understood the use? The AI Act discussion highlights training, internal guidance and AI competence as implementation tasks .

Bottom line

Yes, AI can generally be usable at university and at work in Germany and the EU. But it is safest only when it fits the specific context. For students, the decisive rules are course, exam and university requirements. For employees, the decisive rules are internal approval, data protection, security and the EU AI Act . When AI influences assessments or decisions, the bar can be much higher .