Documentation and labeling of the use of generative AI technologies
in writing and research processes
Various methods are currently being tested in different disciplines to document the use of AI tools. In most disciplines, however, there is still no consensus on how the use of generative AI technologies should be documented and labeled. Here we present various ways in which writers can document and label their use of AI in writing and research processes.
The information on this page is prepared on the basis of student questions.
- Students can use this page as a basis of information to label and document their use of AI applications. The prerequisite for this is that the use of AI corresponds to the framework conditions defined by the teacher.
- Teachers can use this page for guidance and as a decision-making aid to determine how students should document and label the use of generative AI in their work. Teachers may also find these text modules for syllabi helpful when positioning the use of AI by students.
Status: 21.10.2024
How can I indicate in my texts that I have used generative AI technologies? How can I document my use of AI? How can I cite AI tools?
Generative AI technologies can be used for different purposes when writing, e.g. as an aid or as a source. Regardless of what you actually use generative AI for in the research process, you must identify and document its use:
"The verbatim adoption of AI-generated text passages must be identified - analogous to conventional citations - by stating the AI system and specifying the interaction."(Orientation framework for dealing with text-generating AI systems, Vice-Rectorate for Studies and Teaching, 2023, p. 2)
It is not advisable to adopt AI-generated text directly into your own work without checking it. This is because when you submit a text (e.g. an excerpt, a seminar paper or a Master's thesis) for assessment, you as the author are always responsible for the text. It is therefore essential that you critically evaluate AI-generated texts before you continue working with them. In particular, due to the error-prone nature of AI technologies, you should check all information carefully (see question on the limits, weaknesses and limitations of AI) and be sure to include other scientific sources in your research.
Why is the labeling of the sources and aids used in academic writing important?
The overarching aim of labeling the sources and aids used is to make it clear to readers which sources and information your argumentation is based on. On the one hand, it is important that you make the sources used transparent so that they can be verified if necessary. On the other hand, many disciplines also require you to document your research and discovery process. This is often done in a methods chapter, in which the research design is presented and the data and tools used (e.g. statistical software) are explained.
For the documentation of AI use, this means that it should be clear to readers which parts of the text and research-relevant information were generated by an AI and how these were used in the research process.
When labeling the use of AI as an aid or source, the following aspects should always be taken into account:
AI-generated texts are not academic sources
Although AI-generated texts can be based in part on academic sources, they themselves do not have the status of an academic source because they have not undergone the established quality assurance processes (e.g. peer review). When writing, they should therefore be "treated more like the result of a standard Internet search" (University of Basel).
Currently different methods are being tested in various disciplines to document the use of AI tools. In most disciplines, however, there is still no consensus on how the use of generative AI technologies should be documented and labeled. In contrast, there are clear rules for direct and indirect citations from conventional publications (e.g. monographs or articles from specialist journals), which are set out in various citation systems.
Discuss documentation with teachers in the context of their own discipline
In courses, teachers can define specific guidelines for marking in syllabi, rules for tools or academic integrity statements.
It is recommended that you evaluate the advantages and disadvantages of the different documentation options and decide on this basis which of the options is appropriate for your work process. In the event that your instructors do not provide precise guidelines for documentation, you should also discuss the chosen documentation system with them.
Use AI-generated data as a primary source
In some cases, AI-generated products are used as the primary source. In this case, AI-generated outputs are the data basis for answering the research question and the AI application itself is the subject of the research. For example, a sociological study could investigate the extent to which AI-generated texts about professions are characterized by stereotypical gender perceptions.
If you use AI systems as a primary source, you can describe your approach in an appropriate place in your work (e.g. methods chapter) and document the data according to the guidelines of your discipline.
Observe license and terms of use
The license and terms of use of some AI software providers require that the use of the AI application be disclosed.