Thesis Guidelines

These are the “rules of the road” I use with my students. They don’t replace Canvas or official UT/IEM instructions—they help you make good decisions early, before time pressure kicks in.

Timeframes and Page Limits

Use the official rules (Canvas) as the final authority. This toolkit helps you plan and avoid the most common (and expensive) timing mistakes.

  • Start from Timeframes and Page Limits to set a realistic scope.
  • Treat page limits as a budgeting tool: if you add pages in one chapter, you need to remove pages elsewhere.
  • If you are unsure whether your scope fits, bring a one-page outline to a meeting—we can adjust early (cheap) rather than late (expensive).

Confidentiality

If your thesis involves sensitive company information, decide on confidentiality early. It affects what you can show in your thesis, presentation, repository, and even abstract.

If a company asks for full confidentiality, you need to go to the exam committee (do it ASAP; the procedure may take long).

Otherwise, to hide sensitive information, you can, for example:

  1. Write “Company X”, “Product Y”, etc.
  2. Put the confidential information in the appendix and do not publish the appendix.
  3. Remove or generalize identifiers in figures/tables (names, addresses, customer IDs, internal KPIs), and explain what you changed.
  4. Confirm early what is allowed to be shown in the abstract and conclusions (these are often the most visible parts).

Practical advice:

  • Ask the company (in writing) what can appear in the thesis and what can appear in your colloquium slides.
  • Remember: the abstract, conclusions, and figures often get shared the widest. Make these safe by design.
  • If the company requests full confidentiality, start the formal procedure immediately; administrative lead times can be longer than students expect.

Agreements for Theses at Companies

If you are working on your thesis with a company, a formal agreement is required:

  • Three-Party Agreement: This must be signed by you, the university, and the company. Standard forms are available on Canvas.
  • Important: UT supervisors cannot sign these agreements. For help, contact the Internship Coordinator at internshipcoordinator-bms@utwente.nl.

Ethics Approval (When Applicable)

Some projects require ethics approval before you can start the relevant research activities (for instance interviews, surveys, experiments with participants, or working with sensitive personal data).

If you are unsure whether ethics approval applies to your project, check the current Canvas instructions and ask early. The process can include supervisor approval/actions, so don’t do it last minute.

Writing Quality

Before sending drafts for feedback, use the Report Quality Check. It is faster for both of us if you run the checklist first and send a clean draft.

A good submission for feedback usually includes:

  • A complete unit of writing (a full section or chapter draft, not scattered paragraphs).
  • A short note: what changed since last time, what decisions you want to make next, and where you want critique.
  • Any context needed to review (assumptions, data definitions, how you computed results).

Academic Integrity and Similarity Reports

For the greenlight process, you are required to include a thesis similarity report (Canvas provides the mechanism). You are responsible for generating it and sending it to me before the greenlight meeting.

Using AI Tools (UT Policy and Practical Advice)

UT does not have a single “yes/no” rule for every assignment: the allowed use of generative AI is set by the instructor and must be communicated clearly for each assessment (Gertz & Demirtas, 2023). If the course/assignment doesn’t say, ask—don’t guess.

My policy is simple: you may use AI tools to support your work, but you must own the work.

  • You remain responsible for every claim, definition, citation, and number.
  • You should be able to explain your choices and reproduce the steps (methods, code, analysis) without the tool.
  • Treat AI as a drafting assistant, not an authority: verify facts and citations, and rewrite in your own voice.
  • Never paste confidential company data, unpublished results, or personal data into public AI tools (privacy and confidentiality still apply) (University of Twente, n.d.).
TipPlain Language Beats Fancy Language

A useful rule (often attributed to Mark Twain): “Don’t use a five-dollar word when a fifty-cent word will do.”

This matters because many students (and readers) are non-native English speakers. AI tools sometimes “improve” a draft by choosing uncommon words or idioms that make the text harder to read. If you use AI, explicitly ask for clear, simple academic English, then do a quick pass to replace rare words with common ones.

TipDrafting Is Fast; Editing Is the Work

There is an old line often attributed to Blaise Pascal:

“I would have written a shorter letter, but I did not have the time.”

AI tools can help you turn outlines and notes into a readable first draft quickly. But the time you save in drafting should go into revision: tighten structure, remove repetition, check every claim, and simplify the language.

Finally, aim for quality. Avoid “AI slop”1—but note that “human slop” (rushed, vague, citation-free writing) is just as harmful. If the content is rigorous and the writing is clear, I don’t mind how you got to the first draft. Some examiners may be stricter, so the safest default is: follow UT and course rules, and when in doubt, disclose and discuss early (Gertz & Demirtas, 2023).

1 “AI slop” is a term for low-quality, generic, and often misleading AI-generated content posted at scale (Nemeroff, 2025). The broader term “slop” was Merriam-Webster’s Word of the Year for 2025, reflecting how mainstream this concern has become (Merriam-Webster, 2025).

Roles: Supervisor and Examiner

In the UT/IEM thesis process, the second supervisor is often treated primarily as an examiner. In practice, this means:

  • You should not plan your workflow assuming two supervisors will give ongoing, iterative feedback.
  • The examiner’s main involvement is typically at the greenlight meeting and the colloquium, unless specific expertise is needed earlier.

References

University of Twente. (n.d.). Fraud / generative AI (ChatGPT). University of Twente. Retrieved January 14, 2026, from https://www.utwente.nl/en/examination-board/Fraud/
Merriam-Webster. (2025, December 14). Word of the year 2025: slop. Merriam-Webster. Retrieved January 14, 2026, from https://www.merriam-webster.com/wordplay/word-of-the-year
Nemeroff, A. (2025, September 2). What is AI slop? A technologist explains this new and largely unwelcome form of online content. The Conversation. Retrieved January 14, 2026, from https://theconversation.com/what-is-ai-slop-a-technologist-explains-this-new-and-largely-unwelcome-form-of-online-content-256554
Gertz, N., & Demirtas, D. (2023, February 9). ChatGPT policy recommendations. University of Twente. Retrieved January 14, 2026, from https://www.utwente.nl/en/examination-board/Fraud/chatgpt-policy-recommendations-feb23.pdf