Use of artificial intelligence
Different artificial intelligences may provide incorrect information on research topics or even create fictitious sources, which is why the output must always be checked. In addition, some content may contain bias (e.g., discriminatory content), which means the output should be critically assessed. In general, AI outputs are often a good starting point, but they must be critically reviewed and adapted.
AI tools can be applied in different ways. Here are a few examples:
| Area of Application | Description | Example Tools |
| Brainstorming | Ideas for examples, introduction, headings, study design, implications, further research | Perplexity, You, Writsonic, Flow GPT, ChatGPT, Bing |
| Research | References to relevant data and facts, source recommendations | Connected papers, Elicit, Explain papers, Research rabbit |
| Table of Contents | Developing a rough structure for a seminar paper | ChatGPT, Writesonic |
| Text Creation | Formulations, spelling check | ChatGPT, DeepL Write, Grammarly |
| Translations | Translation of literature, translation of qualitative data | DeepL |
In the templates for academic papers as well as in the AI disclosure template, a table is included that must be completed in full when artificial intelligence is used. The specific use of AI and the prompts used must be documented transparently in this table. If the tool used allows the generation of a tracking link (e.g., with ChatGPT), this must also be provided in the last column of the table.