Maintaining independence: What has been my contribution?
If you use AI whilst studying, the key factor is how much of your own work goes into your assignment. A helpful orientation is to distinguish between AI-generated content and AI-assisted work.
This distinction will help you assess
- what counts as your own work,
- what might be problematic,
- and how you should ensure transparency regarding the use of AI
-
AI generates academic content
If sections of text or entire texts are taken directly from an AI tool without being substantially revised, this is not regarded as an independent academic achievement.
In this case, the key decisions regarding content are made by the AI system. Your own contribution is limited to the prompt, which leaves the key decisions to the AI and does not provide sufficient guidance on the content.
Such practices are contrary to the principles of good scientific practice.
-
AI serves as a tool to assist doing your work
For a text generated by AI to be regarded as an original work, there must be a significant human contribution that substantially determines the form of the generated text. In this case, the AI tool serves as an instrument for carrying out the work. This is the case
- if there is a highly sophisticated prompt that has been created by the user themselves and which may even require knowledge relevant to the examination to create,
- if text is generated through the use of highly directive prompts that dictate the key creative decisions rather than leaving them to the AI tool,
- if users independently adapt the generated text as food for thought. (cf. Salden und Leschke 2023)
As an assisting tool, AI complements the academic work process without replacing your own contribution. It is important that you retain control over decisions regarding content. This is the case, for example, when you use AI
- use it to generate ideas or as food for thought,
- to obtain suggestions for structure or organisation that will form the basis for further work (not to be adopted verbatim)
- use it as a tool for reflection, for example to examine arguments,
- or use it to edit your own texts.
Controlling AI in a targeted manner and maintaining independence
The quality of AI-generated output depends to a large extent on how precise a request is. AI tools do not independently respond to academic requirements; rather, they orient themselves to the specifications they receive through prompts.
It is important to define the objective, context and framework conditions clearly, as this will ensure the most effective use of AI as a supportive tool. Through targeted prompting, the user retains control over the working process, determines the direction and depth of responses, and ensures that substantive decisions remain theirs. AI thus provides impulses – the academic contextualisation and further development remain the responsibility of the user.
Prompt tip: How to write prompts to ensure you get credit for your own work
Write prompts in a way that AI does not think or write for you, but rather supports you in your own thinking by:
- Asking for questions, alternatives, feedback or suggestions for structure, not for finished texts or decisions on content. Prompts that do the following have proven effective:
- assigning AI a supporting role (e.g. providing feedback, aiding reflection),
- clearly stating what the AI should not do (e.g. no ready-made arguments),
- making it clear that you want to select, asses and further process AIs output
AI output serves as a starting point and must be subject to review, categorisation and further development.