Skip to Main Content

Referencing and plagiarism

Generative AI

Acknowledging the use of AI and referencing AI

This guidance is intended to help you implement Cranfield University's provisional guidance for the use of generative AI, specifically with regards to referencing generative AI when you have used it. However, please note that generative AI is evolving rapidly and there is not yet general consensus on how to acknowledge and reference it. This guidance will therefore continue to be reviewed and updated.

Considerations for using generative AI in academic work

Before using generative AI, it is essential you ensure that:

  • You know whether or not it is permitted for your assignment or research.
  • You understand the limitations and risks of using generative AI.
  • Your assignment or research remains your own work.

Generative AI can be a useful starting point to gather background information on a topic, but be aware that:

  • Generative AI produces information that may be inaccurate, biased, or outdated.
  • Generative AI is not an original source of information. It reproduces information from a variety of unidentified sources.
  • Generative AI may create fictional quotations and citations.
  • It is always best to refer to original and credible sources of information.

If you do choose to use generative AI tools, you must always:

  • Critically evaluate any output it produces.
  • Carefully check any quotations or citations it creates.
  • Correctly document your use of the tools so that it can be appropriately acknowledged.

Acknowledging the use of generative AI in academic work

The use of generative AI must be acknowledged in an ‘Acknowledgements’ section of any piece of academic work where it has been used as a tool to assist in the process of creating the work.

Minimum requirement to include in acknowledgement:

  • Name and version of the generative AI system used: e.g. ChatGPT-3.5
  • Publisher (company that made the AI system): e.g. OpenAI
  • URL of the AI system.
  • Brief description (single sentence) of context in which the tool was used.

For example: 
I acknowledge the use of ChatGPT 3.5 (Open AI, to summarise my initial notes and to proofread my final draft.

Further requirements may be stipulated by a department, academic programme or individual teaching staff, or for a particular assignment, and must be made clear to students when an assignment is set. Additional requirements may include expanded description in the ‘Acknowledgements’ or ‘Methods’ section, such as:

  • If relevant, the prompt(s) used to generate a response in the AI system.
  • The date the output was generated.
  • The output obtained (e.g. a ‘link to chat’ if ChatGPT, or a compilation of all output generated as an appendix).
  • How the output was changed for use or incorporation into a piece of work (e.g. a tracked-changes document or a descriptive paragraph).

These acknowledgements should not be included in the word count of a piece of work, unless stipulated otherwise for a particular assignment or by a particular academic programme or department. The acknowledgements should either be placed at the beginning or end of the document.

In-text citations and inclusion of AI generated output in a reference list

Some referencing styles suggest that AI systems should be cited in a similar way to other sources, most notably personal communications, but there are issues with citing AI systems:

  • An AI tool cannot be classed as an author – it cannot take responsibility for its work, nor does it generate original ideas but reproduces ideas found elsewhere.
  • One of the primary functions of a reference list is to direct the reader to the original source, which is not possible with AI generated content.

We therefore favour the approach of many academic publishers, which stipulates that AI systems should not be cited as an author, nor included as a source in the reference list.

However, there may be cases where it is appropriate or necessary for you to refer to AI generated output within a piece of work and include it in a reference list, for example:

  • Where the piece of work addresses the topic of generative AI and discussion around outputs.
  • Where there is reference to a formally published output generated by AI.
  • Where it is required by the academic department.
  • Where you have not identified a primary source of the information despite the issues with relying on generative AI as a secondary source of information (which may be considered poor academic practice).

In such cases, the output should be treated as a work with no author, unless specified otherwise by departmental guidelines or the standardised referencing style you are using.

Use of standardised referencing styles

If you are required to use a standard referencing style which has specific rules for citing and referencing AI, the rules of the citation style should be followed, in addition to acknowledging the use of AI as outlined above. This may require the use of in-text citations and reference list entries where the AI tool is listed as an author.

Examples of how generative AI can be cited in a reference list are given below for the styles that are used at Cranfield University.


Open AI. (2023, November 6). ChatGPT (Version gpt-3.5-turbo-1106) [Large language model].


Open AI. ChatGPT [large language model on the Internet]. Version gpt-3.5-turbo-1106. San Francisco: Open AI; 2023 Nov 6 [cited 2023 Nov 22]. Available from:

Further reading

Stokel-Walker, C. (2023) ChatGPT listed as author on research papers. Nature, 613(7945), 620-621.

Cranfield University acknowledges the guidance provided by UCL in producing this information.