
Whether intentional or unintentional: in academia, plagiarism contradicts the principles of scholarly work. Consequences include failed exams, in severe cases even expulsion and failed careers in science and politics.
The most common forms of plagiarism are:
1. Direct plagiarism or full plagiarism:
The verbatim copying of text passages without quotation marks and without citing the source.
2. Mosaic plagiarism or patchwork plagiarism:
The compilation of text fragments from different sources that are inserted into one’s own text without quotation marks and citations.
3. Paraphrase plagiarism:
Rewriting texts of others in one’s own words without proper citation. Even if the wording is changed, the idea or line of thought of the original text remains and must therefore be cited.
4. Structural plagiarism:
The adoption of outlines, structures, or chains of argument from other texts without pointing this out.
5. Idea plagiarism:
The use of ideas, concepts, theories, or research results of others without citing them. Even if no exact text passages are copied, the source of the ideas used must be indicated.

Whether you give sources when adopting a structure or an idea depends on the so-called level of originality:
If you present generally known, undisputed knowledge in your field, usually no citation is required. Sometimes it is hard to determine whether something belongs to this basic knowledge in the field. When in doubt, cite too much rather than too little.
However, if it is an original idea, a specific number, a model or a theory that can be traced back to a person, you must clearly indicate its origin by a citation.
The same applies to adopting a structure. If you take over an innovative outline from academic literature, you can, for example, mention in the introduction that you based your structure on source XY.

You may have read that generative AI does not produce plagiarism, but unique texts. And that is true – these large language models do not simply copy sentences from their vast training data but generate a new, independent text for each prompt according to statistical principles.
As a rule, AI writing tools therefore do not create literal plagiarism. Nevertheless, AI writing tools generate texts that are partially or entirely based on existing content.
Particularly problematic are the following practices:
- If you have scholarly literature rephrased or summarized, you must always cite the source. Otherwise it is plagiarism.
- If you take over generated texts directly from AI tools, you may not recognize that these contents are based on existing works.
- AI tools usually provide no or even false citations for the generated texts, which increases the risk of plagiarism.
To avoid plagiarism when using AI writing tools, you should observe the following strategies:
1. Research carefully:
Science is always based on the prior work of others. It is your academic duty to research thoroughly and take into account what has already been studied on the subject. This also allows you to identify the ideas of others in AI-generated texts.
2. Work independently:
Use AI tools as an aid, not as a substitute for your own thinking and writing. Especially in academic texts, it is often advisable to draft your work yourself and use AI only for revision or supplementation.
3. Provide sources:
Make sure that all information not coming from your own knowledge is correctly cited. Verify content you take from an AI tool like ChatGPT and back it up with sources. To support your literature research, you can also use literature-search AI tools like ConnectedPapers, Consensus, or ResearchRabbit.
4. Reflect on your writing process:
Regularly analyze your writing process and reflect on how and why you use AI tools. This promotes awareness of your own working methods and helps avoid unethical behavior.
5. Use consultations and trainings:
Many universities offer workshops and consultations on avoiding plagiarism. Take advantage of these offers to further develop your skills.
Do I have to list the AI tools I used in the bibliography as a source?
Ultimately, your supervisors decide. Usually it is required to mention AI tools as aids, but not as a source.
Which AI tools are suitable for what?
Take a look at our article on: Using AI tools for academic work.
This article was published in September 2025 and last updated in January 2025.