Yes, as a support tool, if your faculty's guidelines and the instructions for your task allow it, and with a declaration. The Wits AI Policy (G-C-2026-030), which Council approved and which took effect on June 4, 2026, says AI use must be transparently disclosed and that you remain fully responsible for the originality, accuracy, and integrity of your work. Declarations go through a template in the policy's appendix, but that template is marked as a draft sample, to be amended according to faculty guidelines. Check with your faculty which version applies to your thesis, dissertation, or research report.
If your AI statement clearly doesn't cover what you describe elsewhere in your thesis, Esmond points it out.
The policy sets six principles for staff and students. For a thesis, three of them matter most here. You stay fully responsible for the originality, accuracy, and integrity of your work, and using a tool does not excuse academic misconduct. AI should support human intellect without replacing it, and the final judgment, critical interpretation, creative insight, and ethical decision-making stay with you. And AI use must be transparently disclosed and appropriately acknowledged.
The policy also says researchers should describe their AI use in dedicated sections of the final report of a study, so that methods and findings can be reproduced, and that AI cannot be an author on any work. The principles are embedded in faculty-specific guidelines, and the policy tells staff and students to consult those for work in their faculty.
A hypothetical example of a commentOn your declaration you ticked only the box for editing, but your methodology chapter says an AI tool suggested the coding scheme for your interviews. Which part of the form covers that?
A first comment on one part of your own draft, right away and without an account.
Get a free previewOn the form of the declaration, the policy points to its own appendix: “Such declarations are to be made to the University through the comprehensive template provided as in appendix 1 attached hereto.” The closing paragraph of the principles adds that an acknowledgement that you understand the principles should be declared in all formal submissions for assessment, through the same template.
The appendix itself is headed Appendix A and is labeled DRAFT SAMPLE. Under its title it says “(TO BE AMENDED ACCORDING TO FACULTY GUIDELINES)”. So the sample shows the kind of declaration the policy expects, and your faculty may use a different version. Check with your faculty or school office before you sign one.
The sample declaration has three main sections. First, general statements on academic integrity, such as that all sources, human or AI, are cited. Second, statements on specific AI use: that AI-generated content was reviewed and critically assessed by you, acknowledged and cited, and placed in the context of your own work, and that AI use in the design, analysis, writing, or execution of research is disclosed in the relevant sections, for example the methodology or acknowledgements. Third, a choice between two boxes: you did not use AI tools, or you did and you complete a disclosure table.
The disclosure table lists purposes with a tick box, the tool, and a description of how and why you used it. These are the purposes in the sample:
| Purpose in the sample table | Examples the sample gives |
|---|---|
| Idea generation | Research problem, hypothesis |
| Sourcing related work | Summarizing, identifying literature |
| Method and experiment design | None given |
| Data analysis | Coding, visualization, interpretation |
| Theoretical development | Modeling, concept clarification |
| Code development | Writing or testing scripts |
| Presentation | Graphics, visuals, layout |
| Editing | Grammar, clarity, language |
| Writing | Generating, structural suggestions, rephrasing, content suggestions or expansion |
| Citation formatting | Organizing references |
The last line of the sample form names all three: if generative AI tools were an integral part of the experimental design or the direct execution of your research, you confirm that this is outlined in the experimental or methodology chapters of your thesis, dissertation, or research report. The form ends with your signature and the date. The general statements at the top refer to an assignment, essay, proposal, or research report, which is one reason to ask your faculty whether there is a version for postgraduate research.
The policy is to be read with the Wits POPIA Policy and is direct about research data: “Sensitive and / or personal data should not be entered into public AI systems, as this may compromise research participants, or intellectual property.” It adds that using AI on identifiable sensitive or personal data without anonymization or pseudonymization may breach the Protection of Personal Information Act, and that disclosing confidential data can destroy the novelty of your work.
Where AI played a meaningful role in research, the policy also requires an honest record of how and where it was used, which it calls decision provenance. Examples of such declarations are in faculty documents on the Wits AI Policies page.
This example is made up. A master's student in a science faculty submits a research report on water quality. She used an AI tool to fix errors in her analysis scripts and another to tidy her reference list. On her faculty's declaration she ticks code development and citation formatting, names both tools, and describes each use in a sentence. Because the script fixes affected how her data were processed, she also mentions the tool in the methods section, as the sample's statement on AI use in research analysis asks.
Based on the policy and its draft sample. The first item is our suggestion.
Esmond evaluates every assignment before you submit. It never writes it.
Esmond is an AI tool that doesn't write for you. If your course doesn't allow any AI tools, don't use Esmond for that work. If you're unsure, ask your instructor or advisor.
The policy doesn't mention feedback tools, and no row in the sample table names feedback on your own draft. The sample also has a field for other uses. Our suggestion is to tell your supervisor first, to check your faculty's guidelines, and to declare Esmond in that field with a short description, unless your faculty's form says otherwise. The policy also warns against putting personal data into public AI systems, so read the Privacy Notice and remove identifying details of participants before you upload. The University of the Witwatersrand has not reviewed or approved Esmond.
Esmond reads the draft you prepared and flags factual errors, inconsistencies, and blind spots. It adds no text to your thesis, dissertation, or research report. See how Esmond works and the sample report.
Not as printed. The appendix is labeled DRAFT SAMPLE and says it is to be amended according to faculty guidelines. Check with your faculty which version you need to sign.
The sample form asks you to tick one of two boxes, one of which confirms that you did not use AI tools. The policy says an acknowledgement of its principles should be declared in all formal submissions for assessment.
Council approved the policy, reference G-C-2026-030, on June 4, 2026, and it took effect the same day. It is to be read with the Wits POPIA Policy.
The policy says sensitive or personal data should not be entered into public AI systems, and that using identifiable personal data without anonymization or pseudonymization may breach POPIA. Check with your supervisor and your ethics approval.
Read on November 7, 2026.
Get your paper evaluated before you submit