1. Home
  2. About
  3. AI policy
Our AI policy

Responsible use of AI in research

AI tools can save time in research. They can also invent references, expose confidential data and put authorship at risk. This is how we use them in your project, and how we don't.

How we may use AI tools

  • Help with code for data cleaning and analysis. Every script is read, run and checked by our team.
  • Language polishing and formatting. The scientific content comes from the research and from you.
  • Support with literature searching and screening. Every reference is checked against its original source.
  • Summarising guidelines, TORs and donor templates for our own working notes.

What we never do

  • Generate, alter or "fill in" data or results.
  • Cite a reference we have not checked in its original source.
  • Put identifiable or confidential client data into public AI tools.
  • Present AI-generated text as your own intellectual work.
  • Rewrite text to get around plagiarism or AI-writing detectors.
  • Let an AI tool make scientific judgements, or list one as an author.
Our commitments

People stay accountable for every deliverable

A person checks everything

Every deliverable is prepared and checked by a named team member and quality-reviewed by our Principal Consultant. An AI tool is never the final check.

We tell you what we used

If an AI tool helped with any part of your deliverable, your Delivery & Handover Note says which part and how, so you can disclose it as your university, journal or funder requires, for example under ICMJE and COPE guidance.

Your data stays protected

We work only with de-identified data, under a signed NDA. When an AI tool is used with project material, it is set up not to keep or learn from it, and it is never used with identifiable data.

Originality and AI-writing reports

Checked before you submit, explained in plain words

Written academic deliverables come with an originality report covering similarity and an AI-writing check.

  • Our check does not keep your document in any shared database, so your official submission will not later be matched against it.
  • AI-writing scores are estimates, not proof. We explain what your report means, and what to do if a later check raises a question.
  • Your institution's or journal's own check is the one that counts.

Learn it yourself

Ethical use of AI in research is one of the first courses we are preparing at Nebula Academy, for scholars, faculty and research teams.

Join the interest list

This policy applies to all Nebula R&D projects and associates, alongside our Privacy Policy and the Service Agreement & NDA. Last reviewed: November 2026.

Questions about AI in your project?

Ask us in your brief or during the free data discussion. We will tell you exactly how your work will be done.

Chat on WhatsApp