Responsible Use of AI in Research

Use of AI and digital assistance tools may be acceptable where it supports research practice and does not compromise integrity, confidentiality, accountability or originality. This may include:

• Supporting literature discovery, searching and organisation
• Assisting with planning, note-taking or summarising information for further review
• Helping with coding, data processing or workflow support, where outputs are checked by the researcher
• Improving grammar, expression, structure or readability in draft material
• Supporting transcription, translation or formatting, where appropriate safeguards are in place
• Assisting with administrative or teaching activities related to research
• Supporting publication preparation, where use is consistent with publisher requirements and is appropriately acknowledged when needed.

Use of AI and digital assistance tools is not acceptable where it undermines research integrity, obscures responsibility, or creates avoidable risk. This includes:

• Presenting AI-generated content, analysis or interpretation as entirely your own where acknowledgement is required
• Relying on AI outputs without checking their accuracy, completeness or suitability
• Entering confidential, sensitive, personal, proprietary or otherwise protected information into tools that do not provide appropriate safety for that information
• Using AI tools to fabricate, falsify or misrepresent data, sources, results or references
• Using AI in a way that breaches ethics approval conditions, privacy obligations, contractual obligations, copyright, publisher rules or University requirements
• Using AI tools in place of appropriate researcher judgement, expertise or supervision
• Attributing authorship to an AI tool or using AI in a way that misrepresents intellectual contribution.

AI and digital assistance tools can assist research, but they do not replace human responsibility. Researchers, staff and students remain accountable for the integrity of their work and for decisions about how such tools are used.

In all cases, users remain responsible for reviewing outputs, checking accuracy, applying judgement, and ensuring that the final work meets required academic, ethical and professional standards.

Standard authoring tools such as Microsoft Word provide spelling and grammar checks, whilst services such as Grammarly assist with rephrasing and expression. More recent generative AI tools can undertake much more extensive rephrasing or drafting.

Whether rephrasing text moves beyond minor editing assistance, and therefore impacts the integrity of the research output, depends on whether the rephrasing is beyond that acceptable according to the Australian Standards for Editing Practice. These standards limit editing to language and expression (clarity, grammar, spelling), completeness and consistency. Substantial redrafting of whole paragraphs or larger sections, or translating a substantial body of text into English, falls outside those limits.

For research outputs, a distinction can be made between editorial assistance and content assistance offered by more sophisticated digital assistance tools. Editing assistance using AI tools may be acknowledged, but acknowledgement is not required for digital assistance tools that limit edits to minor changes.

There are legitimate uses of generative AI and other such tools in research. The University of Melbourne emphasises that all authors are responsible for ensuring the correctness of published work. If a paper includes text generated by AI then the onus is on the authors to establish that it is correct and sufficient by checking against other sources, and to acknowledge the content assistance provided for specific sections of the work. The use of AI generated text should be cited appropriately; see the University’s guidance on referencing styles for more information.

Several digital assistance tools can help with preparing literature reviews, however, as set out in the Using AI as a graduate researcher page, the University discourages use of generative AI to write literature reviews.

These tools can omit, misrank or even invent material and should not be relied on to understand a field. The University recommends only using them to suggest readings or to shape searches. Authors should always verify outputs against primary sources.

To quote the output of a digital assistance tool that has provided content assistance, it is recommended that:

• Authors verify accuracy and provenance by undertaking the necessary due diligence to confirm references, claims and facts against primary sources
• The material be subject to a plagiarism check, as generative AI output can be a close paraphrase of another source
• The material be subject to a copyright infringement check, as generative AI may produce outputs protected by copyright law
• A footnote can be used instead of an in-text citation, stating the tool used, the version (if available), the date on which it was used, and the exact text of the prompt
• Authors should be explicit as to which material, such as specific passages of text or images, were created or modified by AI tools. Additionally, note that many publishers explicitly prohibit generated images.

A specific form of AI output is transcriptions of spoken text. Such material can only be uploaded to an AI if appropriate permissions or consents are in place. Also, it is critical that the correctness of such transcriptions is thoroughly checked, particularly if they are quoted in publications or used for quantitative analysis. Failure of a transcription tool is not a defence against ethical breaches such as misquoting.

Beyond editing assistance and content assistance for text, note that the use of digital assistance tools to falsify or fabricate (in the sense of ‘inventing in order to deceive’) data is a breach of the University of Melbourne’s Research Integrity and Misconduct Policy and the Australian Code for the Responsible Conduct of Research (the 2018 Code).

Authors should be aware that unacknowledged use of digital assistance tools to generate or alter text that authors imply is their own may constitute a breach of the 2018 Code. Even when the use of digital assistance tools for content assistance is acknowledged, this may impact the degree to which the work is considered the authors’ own (with impacts for GR thesis examination) and may trigger plagiarism concerns.

For coursework subjects, the University has published advice that sets out, as for research writing, that all use of LLMs in work submitted for assessment must be fully acknowledged, and this may have an impact on assessment outcomes.

Researchers should be aware that, when using online tools, text provided as inputs is offered without copyright protection for others to use freely; likewise, any generated text they use may breach copyright.

If you have any questions about the use of generative AI tools in research or have any suggestions for further resources about GenAI and research, please contact Chancellery (Research and Enterprise) at research-enterprise@unimelb.edu.au.