Guidelines for the Use of AI

Principle 1

IfG Impulse welcomes the responsible use of generative AI and AI-supported tools when such use supports scholarly work without replacing human scholarly responsibility.

These guidelines supplement the existing guidelines for authors. The requirements set forth therein regarding scientific soundness, rigor of argumentation, originality, proper citation, and quality of presentation apply regardless of whether AI was used. Authors remain responsible for all content in a submission. This includes, in particular, statements, sources, data, analyses, figures, and conclusions. AI systems cannot assume authorship.

Therefore, what matters is not merely whether AI was used, but for what purpose and what impact it had on the scholarly work.

2. Quick Decision: Green, Yellow, or Red?

GREEN – AI as a Supportive Tool

Permitted, a disclosure is generally not required. Here, AI supports existing human work without generating new scholarly content or conclusions.

Examples:

  • Spelling, grammar, and style corrections
  • Formatting and formal text editing
  • Linguistic refinement of one’s own texts
  • Assistance with search terms or database queries
  • Technical assistance with standard software
  • Improvement of readability without altering the content

The following applies: The authors review the results and are responsible for the final version.

YELLOW – AI influences content or scholarly decisions

Permitted, provided that its use is critically evaluated and transparently disclosed. The yellow category applies whenever AI goes beyond purely technical or linguistic support and contributes to content or influences decisions.

This includes, for example:

  • Ideation and structuring of arguments
  • Summaries of scientific literature
  • Extensive suggestions for phrasing or revision
  • Translations of scientific content
  • Suggestions regarding research design or methodology
  • Generation or revision of code
  • Support with data preparation or data analysis
  • Support with qualitative coding or interpretation
  • Generation of research materials or stimuli
  • Creation of illustrations or visualizations
  • Use of synthetic data, personas, or AI-generated responses

The following applies: AI-generated content must be fact-checked and may not be adopted without verification. Sources, calculations, and statements must be verified using appropriate original sources or methods.

If AI is an integral part of the research design, data collection, or analysis, its use must also be documented in the methods section in such a way that the procedure is transparent. This includes the system or model used, the manner of use, and the procedures for verifying the results.

RED - AI replaces scholarly responsibility or violates scholarly integrity

Not permitted.

This includes, in particular:

  • fictitious or unverified bibliographic references
  • fictitious, falsified, or AI-generated data presented as empirical
  • unverified adoption of facts, analyses, or results
  • generation of central arguments or conclusions that are presented as the author’s own scholarly work
  • Concealment of substantial AI use
  • Use of AI-generated content in a manner that results in plagiarism or inaccurate attribution
  • Input of confidential, personal, or non-disclosable data into AI systems not authorized for such use
  • AI-generated or manipulated figures that convey a false impression of empirical findings or real-world conditions
  • Delegating the scholarly evaluation of a manuscript to an AI system

3. Literature, Facts, and Original Scholarly Contribution

The existing requirements of IfG Impulse regarding literature and sources apply without restriction to AI-supported work as well. Generative AI is generally not a substitute for a scientific source. If AI is used for literature searches or to summarize literature, the original sources used must be verified independently. This applies in particular to authors, titles, publication dates, DOIs, and specific references. The responsibility for the theoretical foundation and the rigor of the argument lies with the authors. AI can support this work but cannot replace it.

An exception applies when an AI system or its output is itself the subject of the academic investigation. In this case, the system used must be appropriately documented.

4. Data, Data Protection, and Illustrations

When using AI, the existing requirements of IfG Impulse regarding data protection, copyrights, and third-party rights must be observed. Personal data, confidential company information, unpublished research data, or other protected content must not be entered into external AI systems unless there is an appropriate legal and institutional basis for doing so. The following applies to AI-generated figures and visualizations as it does to other submitted materials: The authors are responsible for ensuring that the necessary rights have been obtained and that the presentation is factually accurate. AI-generated representations must not give the impression of empirical evidence if none exists.

5. Disclosure of AI Use

Disclosure is required if AI has substantially contributed to the creation of scientific content or influenced research decisions (YELLOW). The statement should briefly indicate which AI system was used, for what purpose, and to what extent.

Example:

“In preparing this article, [system/model] was used to structure individual text sections and to revise the language. All suggestions were reviewed and revised by the authors. References and statements of fact were verified against the original sources.”

If AI was used as part of the research methodology, its use must also be described in the Methods section. Purely technical or linguistic assistance within the “green” category does not need to be disclosed.

6. AI in the Peer Review and Editorial Process

The principles of confidential and independent peer review applicable to IfG Impulse remain unchanged by the use of AI. Reviewers and editorial staff members may not upload unpublished manuscripts or parts thereof to publicly accessible AI systems or those not institutionally authorized. AI may be used to provide permissible technical or linguistic support, but it must not replace the independent substantive review or the judgment of reviewers or editors. Responsibility for peer reviews and editorial decisions remains entirely with the individuals involved.

7. When in doubt: Transparency over concealment

The development of generative AI is advancing rapidly. Not every application can therefore be clearly assigned to a single example in these guidelines. Three questions serve as a guide:

Did the AI merely provide technical support for my own work?

--> GREEN: Use is generally permissible.

Did the AI generate content or influence my scientific decisions?

--> YELLOW: Review, make it transparent, and, if necessary, document the methodology.

Would the use compromise scientific responsibility, verifiability, confidentiality, or integrity?

--> RED: Use not permitted.

In cases of doubt, IfG Impulse recommends making the use transparent and consulting with the editorial board before submission.

As of September 2026. These guidelines are reviewed regularly in light of the dynamic development of AI technologies and scientific publication standards.