Key points at a glance
- Check your university’s policy before deciding whether AI-generated code is allowed or must be disclosed.
- Separate your own implementation, adapted code, generated suggestions and external libraries.
- Record the tool, date, relevant prompt or interaction and the changes you made.
- Citing an AI tool does not replace citations for libraries, algorithms, datasets or original sources.
Start with the academic policy
Rules for generative AI differ by university, programme and assignment. Some policies permit brainstorming but restrict generated code; others require a declaration or an appendix. Read the current policy and ask your supervisor when the wording is unclear.
Your declaration should make the role of the tool understandable without claiming that a model is the author of your scientific argument.
Distinguish the types of code contribution
Document whether you wrote the code yourself, adapted a generated suggestion, accepted a generated block after testing or used an external library. These cases have different implications for authorship, licensing and reproducibility.
Keep the final code, test results and relevant version information. Review every generated line for security, correctness, dependency and licence concerns.
What to record for an AI-assisted contribution
A useful record names the tool and version, date of use, purpose, relevant prompt or interaction and the material changes you made. Store this information in the format your department requests, such as an appendix or methodology note.
Cite the original documentation and repository for any library or framework. An AI transcript cannot prove the provenance of an algorithm or dependency by itself.
Explain your own verification
State how you tested, reviewed and adapted the code. Include meaningful tests or evaluation criteria so another researcher can understand what was accepted and what was rejected.
Do not paste a long transcript into the thesis without context. A concise, reproducible record is more useful than an unstructured export.
Final submission context
Formal details can feel like a separate writing task, but they matter just as much for a printed submission. Anything missing, misplaced or inconsistently formatted in the document will also appear in the bound copy.
Use this guide together with your cover page, table of contents, page numbers, source notes and appendices. Prepare the final PDF for printing and binding only after the complete file has been checked.
Practical check before PDF export
- Do headings, chapter structure and page numbers match?
- Are sources, figures and tables included completely?
- Are required elements such as declarations, appendices or the cover page included where required?
- Did you open and check the final PDF after exporting it?
Checklist
- The current university policy was checked.
- The tool, version, date and purpose are recorded.
- Generated, adapted, self-written and third-party code are distinguished.
- Libraries, algorithms and datasets have their own citations.
- Generated code was tested for correctness and security.
- The disclosure format matches the department’s requirement.
Common mistakes
- Assuming a tool name alone is a sufficient citation.
- Presenting generated code as entirely self-written.
- Ignoring licences and dependencies in generated snippets.
- Sharing confidential source code or personal data in a prompt.
- Failing to record the version and date of the interaction.
Prepare the documented thesis for printing
After your code provenance and AI disclosure are checked, prepare the final PDF for printing and binding.
View thesis printing optionsRelated content
Explore related guidance on thesis structure, sources, formatting and final submission.
Frequently asked questions
Can I cite ChatGPT as the author of my code?
Usually you should document the tool’s contribution rather than assign it authorship. Follow your university’s prescribed disclosure and citation format.
Do I need to cite every generated line?
The exact rule depends on your policy. You should at least make the tool’s role, relevant interaction and your verification process transparent.
Do library citations still matter when AI suggested the code?
Yes. Cite the library, documentation, algorithm or dataset independently of how you discovered or drafted the code.
Where should I document the prompts?
Use the location required by your department, such as an appendix, methodology section or separate declaration.