Information on the Use of LLM/AI

Working Group Modern History and Historical Migration Research
Prof. Dr. Christoph A. Rass
[IMIS] [SFB1604] [HistOS]

Recommendations for dealing with artificial intelligence (AI)

[For courses and qualification theses in the area of the Chair of Modern History and Historical Migration Research]

Preliminary remark

The Chair of Modern History and Historical Migration Research at Osnabrück University is committed to the responsible and transparent use of artificial intelligence (AI) tools in academic work. These recommendations aim to exploit the potential of AI technologies for historical studies while upholding the principles of good research practice, critical source work and academic integrity.

1 Principles for the use of AI

1.1 Critical reflection and personal responsibility

AI tools must always be used in a critically reflective and responsible manner. AI serves to support your own intellectual performance, not to replace it. The responsibility for all content of a scientific paper lies exclusively with the author.

1.2 Transparency and traceability

  • Every use of AI tools must be fully documented and presented in a way that is comprehensible to third parties.
  • This serves to maintain scientific standards and enables a fair assessment of independent performance.

1.3 Subject-specific features

  • In history, the focus is on source work, historical contextualization and methodological reflection.
  • AI tools can support these core competencies, but never replace them.

2 Permitted areas of application

2.1 Supporting functions

AI tools may be used for

  • Brainstorming and idea development: Initial suggestions for questions or outline approaches
  • Literature research impulses: Suggestions for search terms or thematic focuses
  • Linguistic optimization: Improvement of style, grammar and readability in self-written texts
  • Structuring aids: Support with the logical organization of arguments
  • Translation support: help with foreign-language sources (with critical review)
  • Data organization: Systematization of larger source corpora or bibliographic data

2.2 Research support

For extensive research projects, AI tools can be used for

  • Pattern analysis in source material: identification of recurring themes or terms
  • Transcription support: digitization of handwritten sources (with post-processing)
  • Chronological sorting: organization of complex sequences of events

3 Areas of application not permitted

3.1 Core services of scientific work

AI may not be used for:

  • Development of the research question: The central question must be developed independently
  • Historical interpretation: Interpretation and classification of sources and events
  • Argumentation and conclusions: The scientific argumentation must represent own intellectual achievement
  • Evaluation of sources: Source criticism and evaluation are indispensable historical skills
  • Generation of entire text passages: Entire paragraphs or chapters must not be AI-generated

3.2 Methodological core areas

The following areas must be worked on independently:

  • Methodological reflection: justification of the chosen approach
  • Contextualization: Classification in historical contexts
  • Comparative analysis: Evaluation of different approaches to interpretation

4. documentation requirements

4.1 Full disclosure

The following must be stated each time AI tools are used

  • Name and version of the AI system used (e.g. "ChatGPT-4", "Claude 3.5", "DeepL")
  • Specific work step for which AI was used
  • Type of interaction (prompts used, iterative refinements)
  • Concrete contribution of the AI to the work result
  • Critical classification of the benefits and limitations

4.2 Practical implementation of the documentation

Documentation can take the form of

  • Footnotes for selective AI use
  • Method chapters for more extensive use
  • Appendix with a detailed list of all AI interactions
  • Declaration of independence with specific AI information

4.3 Minimum information in the documentation

Example of appropriate documentation: "ChatGPT-4 (November 2024) was used for the linguistic revision of chapter 3.2. The self-written text was entered with the prompt 'Please check this text for grammar and style'. The suggestions were critically reviewed and partially adopted without changing the substantive argumentation."

5 Quality Assurance and responsibility

5.1 Fact checking and source criticism

All AI-generated or AI-supported content must be carefully checked for

  • Factual accuracy: comparison with primary and secondary sources
  • Historical plausibility: Placement in the historical context
  • Source references: Verification of all references suggested by the AI
  • Possible biases: Critical evaluation for systematic bias

5.2 Special caution with historical data

AI systems are prone to "hallucinations" with:

  • Specific historical data and facts
  • Quotations and direct sources
  • Event chronologies and personal data
  • Bibliographic data

All historical information must therefore be supported by primary or verified secondary sources.

6 Legal and ethical aspects

6.1 Copyright and plagiarism

  • AI-generated texts can unknowingly reproduce copyrighted content
  • All AI outputs must be checked for possible plagiarism
  • If in doubt, a plagiarism check must be carried out

6.2 Data protection and confidentiality

  • Do not enter any confidential or personal data into AI systems
  • Exercise particular caution when working with sensitive historical sources (e.g. interviews with contemporary witnesses)
  • Observearchive regulations and data protection requirements

6.3 Scientific honesty

  • AI use must not be concealed
  • Exaggerating one's own performance by concealing AI support is unscientific
  • In case of uncertainties regarding the appropriateness of AI use: consult with supervisors

7 Specific instructions for different forms of work

7.1 Homework and essays

  • AI use is permitted with transparent documentation
  • The focus must be on independent argumentation
  • Declaration of independence with AI documentation required

7.2 Bachelor's and Master's theses

  • Comprehensive documentation of the use of AI in a separate methods section
  • Critical reflection on the limitations and benefits of the AI tools used
  • Comprehensible presentation of the independent scientific achievement

7.3 Papers and presentations

  • AI-supported visualizations are permitted (with identification)
  • Content and argumentation must be developed independently
  • For AI-generated illustrations or graphics: appropriate labeling

8 Step-by-step instructions for AI use

8.1 During editing

  • Before using AI: Clear definition of the desired area of support
  • During use: documentation of all interactions
  • After use: Critical review and verification of all outputs
  • During transcription: Transparent integration into the work
  • Before submission: Complete documentation in declaration of independence

8.1 Checklist for submission

  • [ ] All AI assignments documented?
  • [ ] Historical facts verified by independent sources?
  • [ ] Declaration of independence provided with AI details?
  • [ ] Critical reflection on the use of AI carried out?
  • [ ] Own contribution clearly recognizable?

These recommendations are regularly revised and adapted to the rapidly evolving AI landscape.

Current versions are always available on the Chair's website.

At the same time, these recommendations are intended to be a living document that is continuously developed together with students and teaching staff. Feedback and suggestions for improvement are expressly welcome.

Last update: [June 2025]