Generative Artificial Intelligence for PGR Students

Background:

Generative AI resources and applications are expanding at an astonishing rate.  The tools are incredibly powerful and flexible.  Most researchers are exploring how this technology can contribute to their research, but this is also revealing concerns about ill-considered application of it.  As with all new technologies users should take care in assessing the usefulness and potential pitfalls of GenAI.  PGR students in materials are encouraged to engage with these tools, but to do so responsibly and with careful reflection on safety, reliability and integrity of that use.   

 

Resources:

A useful set of resources on GenAI are available in Oxford through the AI Competency Centre.  This includes a description of GenAI tools available with Oxford single sign on that have gone through Third Party Security Assessment (TPSA) process and include enterprise agreements for robust data protection.  The AI Competency Centre also have a list of training resources targeted at a wide range of experience levels. 

 

Policy and Guidance for Summative Assessment

There are particular aspects that are important to consider for summative assessments (Transfer, Confirmation, and examination).  The University has provided policy on the use of GenAI in summative assessment, and this has been usefully supplemented by guidance for students in the MPLS division.  All PGR students must comply with this policy and guidance.   

The MPLS guidance requires students to include a statement on their use of Generative AI in their documents submitted for transfer and confirmation of status as well as in their final submitted thesis.  In Materials this should be accomplished as follows: 

 

Transfer of Status: 

All DPhil and MSc(Res) students in materials must include a description of their use of generative AI in their Research Project Report.  This is detailed in the Notes of Guidance on Research Degrees in Materials for the relevant degree as updated from May 2026.  Students must provide a declaration that their generative AI use complies with university and divisional policy and guidance, and statement of which AI tools have been used and for what specific purpose in their work.  As a minimum this should for each different usage case list the name, version and publisher of the generative AI tool.  Students are encouraged to also provide a brief reflection on any steps taken to protect IP, ensure data integrity, and testing reliability/robustness of outputs. 

 

Confirmation of Status: 

All DPhil students in materials must include a description of their use of generative AI as a milestone document supporting their application for confirmation of status.  This is detailed in the Notes of Guidance on Research Degrees in Materials as updated from May 2026.  Students must provide a declaration that their generative AI use complies with university and divisional policy and guidance, and statement of which AI tools have been used and for what specific purpose in their work.  As a minimum this should for each different usage case list the name, version and publisher of the generative AI tool.  Students are encouraged to also provide a brief reflection on any steps taken to protect IP, ensure data integrity, and testing reliability/robustness of outputs. 

 

Final Thesis: 

All DPhil and MSc(Res) students in materials must include in their thesis a statement on their use of generative AI.  This statement should be placed immediately after the abstract and must include a formal declaration that any generative AI use complies with university, divisional and departmental guidance.  The statement must also document where and how generative AI has been used in preparation of the thesis and summarise how specific uses of generative AI are referenced in the text of the thesis (this could simply amount to reference to a scientifically accepted standard in place at the time of submission of the thesis).  As a minimum the statement should for each different usage case list the name, version and publisher of the generative AI tool.  Students are encouraged to also provide a brief reflection on any steps taken to protect IP, ensure data integrity, and testing reliability/robustness of outputs. 

 

 

 

 

An Example - Statement on use of generative AI 

I declare that my use of generative AI tools complies with applicable University, divisional, and departmental policy and guidance. 

Generative AI tools used: 

  • Claude 4 Sonnet (Anthropic) — used to discuss possible structures for the methodology chapter and identify areas where additional citations might be required.  

  • Elicit (Version current as of May 2026, Ought) — used to assist in locating potentially relevant literature for screening.  

  • ChatGPT (GPT-5.5, OpenAI) — used to compare my notes from papers I had read so as to identify areas of consensus and areas with diverging results or interpretations.    

  • ChatGPT (GPT-5.5, OpenAI) — used to assist with translating legacy scientific code written in Fortran into Python. 

Specific limitations on AI use: 

  • AI tools were not used to fabricate references, generate experimental data, conduct statistical analysis autonomously, or write unrevised dissertation sections.  

  • All cited literature was independently accessed and verified by the author.  

Reliability and robustness checks: 

  • AI-generated summaries were cross-checked against the original papers.  

  • Suggested references were verified manually for accuracy and relevance.  

  • Code was reviewed, edited, and tested by me before inclusion in the project. I independently verified that the translated Python implementation reproduced the expected numerical outputs of the original Fortran code using benchmark datasets and regression tests.  

Data protection and IP considerations: 

  • No confidential, commercially sensitive, or participant-identifiable data were entered into public AI systems.  

  • Draft manuscripts shared with AI tools excluded unpublished proprietary content. 

 

 

AJW 
28/05/2026