Use of AI in Research Guidelines | University of Technology Sydney

The Use of AI in Research Guidelines (the guidelines) have been developed to help researchers make decisions about the use of Generative AI (GenAI) in their research.

The guidelines are structured to give guidance at each stage of a research project. The guidelines refer to the:

principles of responsible research conduct (code principles column) published in the Australian Code for the Responsible Conduct of Research, and
risk types (risk consequences column) when using AI (as published at the Massachusetts Institute of Technology: AI Risk Repository).

Before researchers use GenAI they should consider that the use of such tools could impact research integrity principles.

Researchers need to ensure that their research complies with the Australian Code, the Research Policy and legislative requirements including intellectual property and privacy.

1. Before you begin

Initial planning and understanding of research integrity principles.

Code principles	Guideline for responsible use	Risk consequences
Respect, Recognition	Classify data What is the classification of the data you will be using? Identify data classification by using the Information Security Classification Standard (available at Records and archives hub: Information security (SharePoint)).	Privacy and security
Transparency, Accountability	Seek advice Have you spoken with your supervisor and experts in the use of AI in research to get advice before using GenAI?	Reputational
Respect, Recognition, Fairness	Maintain confidentiality Are you breaching confidentiality (ethical, cultural or commercial) or licences or agreements by uploading the data into a GenAI tool? Check for restrictions in contracts, agreements or licences.	Privacy and security Intellectual property
Accountability, Respect, Recognition	Maintain cybersecurity If you are considering using a GenAI tool that will have access to UTS IT or network resources, or be deployed on a UTS device, you must follow relevant IT policies, including the Information Security Policy, the Acceptable Use of Information Technology Resources Policy and the Data Governance Policy.	Privacy and security Malicious actors
Transparency, Accountability	Comply with licences and contracts Before you enter any content into a GenAI tool, you need to check the licence or terms and conditions to ensure the content can be used with the tool. Some content may be restricted by licences, agreements and copyright protections and cannot be used in GenAI tools. This may include content from the library (such as journal articles and books) or university-owned IP, including trade secrets or material generated during research activities.	Reputational Loss of revenue Intellectual property
Respect, Recognition, Rigour	Use of synthetic personas Care must be taken when using synthetic personas in research.	Human-computer interaction
Transparency, Accountability	Document use of GenAI Ensure that you document your use of GenAI (refer Definitions) throughout the project.	Human-computer interaction
Respect	Consider resource use GenAI tools have large environmental impacts, ranging from high energy consumption and water use to producing high volumes of electronic waste. Consideration of these factors is recommended when determining use of these tools.	Socioeconomic and environmental
Respect, Recognition	Consider ethics Have you thought about your ethics application, and including the use of GenAI in the consent forms, as well as the participant information sheet (PIS) if you plan on uploading participants’ data into any GenAI tool?	Socioeconomic and environmental

2. Selecting tools

Choosing appropriate GenAI tools for research.

Code principles	Guideline for responsible use	Risk consequences
Rigour	Choose the right tool What are the limitations of the tool and how will this impact your research?	AI system safety, failures and limitations
Rigour	Consider bias Have you thought about how biases might impact the use of the tool?	Discrimination and toxicity
Accountability, Rigour	Assess if the tool meets norms and expectations Choose tools that align with research goals and are fit for purpose, aligning with disciplinary norms and expectations for research methodology and validation.	AI system safety, failures and limitations
Transparency, Respect, Recognition	Consider the impact of data classification on tool selection Is the tool you are using suitable for the type of data you are using? Make sure you appropriately classify the data (check the Information Security Classification Standard and the Information Security Classification User Handling Guide (available at Records and archives hub: Information security (SharePoint)).	Privacy and security
Respect, Accountability	Consider privacy policy of the tool and handling of personal information Have you read the data privacy policies of the tool?	Privacy and security
Rigour, Accountability	Consider how and where data is stored and destroyed How does the tool handle data? Where is the data stored (data sovereignty)? Does the platform outline how long data will be retained and the processes for deletion and purging of data (if dealing with human participant data)?	Privacy and security
Respect, Recognition, Fairness	Consider control of IP and terms of use Be careful about uploading any content into tools. If you want to upload data or other materials, are you aware that you may be at risk of losing control of your own or the university's IP or breaching someone else's IP? Look at the terms of use of any tool you want to use including provisions regarding ownership of and licences to IP. Avoid tools that do not provide clear terms of use.	Privacy and security Intellectual property
Rigour, Transparency	Consider if the tool will access UTS devices, IT resources or network Will the tool have access to any other files on your device, or will the terms of use grant access to any metadata or other information beyond what is directly uploaded? If you are considering using a GenAI tool which will have access to UTS IT or network resources, you must follow relevant IT policies, including the Information Security Policy, the Acceptable Use of Information Technology Resources Policy and the Data Governance Policy.	Privacy and security
Rigour, Respect, Recognition	Consider if data sharing agreements, licensing or contracts permit the use of the GenAI tool Refer to any licences, sharing agreements and/or contracts relating to your project and data to ensure that the use of GenAI tools is permitted, that data can be uploaded to the relevant GenAI tool, and that access to data and outputs is appropriately controlled. Consider if the proposed GenAI tool use is consistent with any research data management plans (refer Research Data Management Procedure).	Privacy and security

3. Project design and codesign

Using GenAI for brainstorming and project design.

Code principles	Guideline for responsible use	Risk consequences
Honesty, Transparency	Document use of GenAI Ensure that you document your use of GenAI throughout the project.	Socioeconomic and environmental
Rigour, Accountability	Address bias in AI outputs Have you thought about how biases might impact the use of the tool? What are the limitations of the tool and how will this impact your research? Choose tools that align with research goals and are fit for purpose, aligning with disciplinary norms and expectations for research methodology and validation.	Discrimination and toxicity
Recognition, Rigour, Accountability	Indigenous research participants, partners and communities Any project using Indigenous Data, Indigenous Knowledges or projects collaborating with Indigenous communities or research participants must be done in accordance with the AIATSIS Code of Ethics for Aboriginal and Torres Strait Islander Research and CARE (Collective Benefit, Authority to Control, Responsible, Ethics) Principles for Indigenous Data Governance. Use of Indigenous Cultural and Intellectual Property (ICIP) with GenAI tools must align with relevant laws, cultural protocols and principles relating to access, use, attribution and benefits as well as any agreements with communities or rights holders. This needs to be viewed holistically in alignment and relation to other guidelines under the respect and recognition code principles.	Discrimination and toxicity

Code principles

Guideline for responsible use

Risk consequences

Honesty, Transparency

Document use of GenAI

Ensure that you document your use of GenAI throughout the project.

Socioeconomic and environmental

Rigour, Accountability

Address bias in AI outputs

Have you thought about how biases might impact the use of the tool? What are the limitations of the tool and how will this impact your research? Choose tools that align with research goals and are fit for purpose, aligning with disciplinary norms and expectations for research methodology and validation.

Discrimination and toxicity

Recognition, Rigour, Accountability

Indigenous research participants, partners and communities

Any project using Indigenous Data, Indigenous Knowledges or projects collaborating with Indigenous communities or research participants must be done in accordance with the AIATSIS Code of Ethics for Aboriginal and Torres Strait Islander Research and CARE (Collective Benefit, Authority to Control, Responsible, Ethics) Principles for Indigenous Data Governance.

Use of Indigenous Cultural and Intellectual Property (ICIP) with GenAI tools must align with relevant laws, cultural protocols and principles relating to access, use, attribution and benefits as well as any agreements with communities or rights holders.

This needs to be viewed holistically in alignment and relation to other guidelines under the respect and recognition code principles.

Discrimination and toxicity

4. Applying for funding

Writing funding applications with the help of GenAI.

Code principles	Guideline for responsible use	Risk consequences
Honesty, Rigour	Check rules for using GenAI to write grant applications Have you checked the funding agency rules and process regarding the use of GenAI in writing grant applications? Follow the process and guidelines of the relevant funding agency.	Socioeconomic and environmental
Rigour, Accountability	Certify accuracy of information If the use of GenAI is permitted by the funding agency for grant applications, ensure that you certify the accuracy and accountability for the information. You are ultimately responsible for authorship and content of the application.	Socioeconomic and environmental
Fairness, Accountability	Check rules before using GenAI to review grant applications Have you checked the funding agency rules and process regarding the use of GenAI for reviewing grant applications? Follow the process and guidelines of the relevant funding agency. The use of GenAI for sensitive or proprietary information, for example grant applications, could compromise the confidentiality of the material and breach agreements.	Socioeconomic and environmental

Code principles

Guideline for responsible use

Risk consequences

Honesty, Rigour

Check rules for using GenAI to write grant applications

Have you checked the funding agency rules and process regarding the use of GenAI in writing grant applications? Follow the process and guidelines of the relevant funding agency.

Socioeconomic and environmental

Rigour, Accountability

Certify accuracy of information

If the use of GenAI is permitted by the funding agency for grant applications, ensure that you certify the accuracy and accountability for the information. You are ultimately responsible for authorship and content of the application.

Socioeconomic and environmental

Fairness, Accountability

Check rules before using GenAI to review grant applications

Have you checked the funding agency rules and process regarding the use of GenAI for reviewing grant applications?

Follow the process and guidelines of the relevant funding agency.

The use of GenAI for sensitive or proprietary information, for example grant applications, could compromise the confidentiality of the material and breach agreements.

Socioeconomic and environmental

5. Reviewing the literature

Conducting literature reviews using GenAI.

Code principles	Guideline for responsible use	Risk consequences
Rigour, Transparency	Appraise and validate Appraise GenAI output for potential biases and inaccuracies by checking outputs for completeness and accuracy. Comparison with traditional bibliographic databases are recommended to ensure reviews are comprehensive.	AI system safety, failures and limitations
Transparency, Accountability	Document use of GenAI Ensure that you document your use of GenAI in literature reviews.	Socioeconomic and environmental
Fairness, Accountability	Comply with licences and contracts Are you breaching terms of any licences or agreements by uploading the materials (books, articles, reports) into a GenAI tool? Check publisher agreements or licences for terms and conditions before uploading such materials.	Socioeconomic and environmental
Honesty, Transparency	Consider bias and transparency Special considerations apply when using GenAI in systematic reviews (refer Definitions).	Discrimination and toxicity

6. Collecting, generating and analysing data

Using GenAI for data collection, generation and analysis.

Code principles	Guideline for responsible use	Risk consequences
Honesty, Transparency	Document use of GenAI Ensure that you capture your use of GenAI in generating or analysing data.	Socioeconomic and environmental
Respect, Fairness	Maintain confidentiality Are you breaching confidentiality (ethical, cultural or commercial) or licences or agreements by uploading the data into a GenAI tool? Check for restrictions in contracts, agreements or licences.	Privacy and security
Recognition, Rigour, Accountability	Indigenous Cultural and Intellectual Property, Indigenous data sovereignty and stewardship, Free prior informed consent and Indigenous participants Any project using Indigenous data, Indigenous Knowledges or projects collaborating with Indigenous communities or research participants must be done in accordance with the AIATSIS Code of Ethics for Aboriginal and Torres Strait Islander Research and CARE (Collective Benefit, Authority to Control, Responsible, Ethics) Principles for Indigenous Data Governance. Use of ICIP with GenAI tools must align with relevant laws, cultural protocols and principles relating to access, use, attribution and benefits as well as any agreements with communities or rights holders. This needs to be viewed holistically in alignment and relation to other guidelines under the respect and recognition code principles.	Discrimination and toxicity
Respect, Recognition	Consider ethics Have you thought about your ethics application and including the use of GenAI in the consent forms, as well as the participant information sheet (PIS) if you plan on uploading participants’ data into any GenAI tool?	Discrimination and toxicity
Respect, Recognition	Respect your research participants If using GenAI to analyse, summarise or distil research participant data (e.g. surveys/questionnaires, interviews or focus groups), there must be an approach to maintain the integrity of the participants’ responses to ensure they are not altered, embellished or diluted. Any use of GenAI on participant responses should be transparently recorded and reported.	Misinformation
Rigour, Accountability	Consider bias Have you thought about how biases might impact the use of the tool? Are there measures in place to mitigate any biases in outputs?	Discrimination and toxicity
Respect, Recognition	Use of synthetic personas Care must be taken when using synthetic personas in research.	Human-computer interaction
Rigour	Verify and validate How will you verify the accuracy and reliability of data generated or analysed by a GenAI tool?	Misinformation

7. Writing up

Writing research outputs such as journal articles, monographs and blog posts.

Code principles	Guideline for responsible use	Risk consequences
Honesty, Accountability	Ensure human authorship Researchers are ultimately responsible for the quality and accuracy of research outputs. Refer to and ensure compliance with the Authorship Procedure.	Human-computer interaction
Honesty, Transparency	Document use of GenAI Ensure that you capture your use of GenAI when writing research papers. Have you used it to refine your writing, or for copyediting and proofreading?	Human-computer interaction
Rigour, Accountability	Verify and validate How are you verifying the accuracy and reliability of any written content generated by a GenAI tool? Are there measures in place to mitigate any biases or inaccuracies in the outputs?	Misinformation
Fairness, Accountability	Ensure attribution and accuracy How are you verifying that content has not been plagiarised or hallucinated? You must verify the existence of all citations used in GenAI writing and ensure they are properly attributed.	Misinformation

8. Publishing your research

Disseminating research findings through publications.

Code principles	Guideline for responsible use	Risk consequences
Honesty, Accountability	Ensure correct verification and disclosure Do not submit GenAI content without proper verification and disclosure. Also refer to Writing up.	Socioeconomic and environmental
Accountability, Transparency	Consider publisher / journal requirements regarding GenAI Have you read the journal/publisher guidelines for who you would like to publish with and what GenAI use they allow? Different journals/publishers have different requirements. Follow guidelines that allow the greatest number of publication options when using and documenting GenAI tools.	Socioeconomic and environmental
Honesty, Rigour	Consider how the work you create may be used by the publisher to train other GenAI models Have you read the journal’s policy about selling or providing your published work as data that can be used for training GenAI models?	Privacy and security

Code principles

Guideline for responsible use

Risk consequences

Honesty, Accountability

Ensure correct verification and disclosure

Do not submit GenAI content without proper verification and disclosure. Also refer to Writing up.

Socioeconomic and environmental

Accountability, Transparency

Consider publisher / journal requirements regarding GenAI

Have you read the journal/publisher guidelines for who you would like to publish with and what GenAI use they allow?

Different journals/publishers have different requirements. Follow guidelines that allow the greatest number of publication options when using and documenting GenAI tools.

Socioeconomic and environmental

Honesty, Rigour

Consider how the work you create may be used by the publisher to train other GenAI models

Have you read the journal’s policy about selling or providing your published work as data that can be used for training GenAI models?

Privacy and security

9. Peer reviewing and editing research

Reviewing and editing research using GenAI tools.

Code principles	Guideline for responsible use	Risk consequences
Fairness, Accountability	Avoid using GenAI for peer review GenAI is not suitable for peer review. The ARC explicitly prohibits the use of GenAI in peer review. GenAI does not have the same disciplinary knowledge and experience as a peer would, nor does it have the discretionary and contextual expertise to be able to apply that knowledge to a piece of novel research.	Reputational risk Intellectual property Loss of revenue
Fairness, Accountability	Breach of confidentiality The use of a GenAI tool for peer review likely breaches the confidentiality of the author(s) and agreements entered into with the journal/publisher.	Privacy and security

Code principles

Guideline for responsible use

Risk consequences

Fairness, Accountability

Avoid using GenAI for peer review

GenAI is not suitable for peer review. The ARC explicitly prohibits the use of GenAI in peer review.

GenAI does not have the same disciplinary knowledge and experience as a peer would, nor does it have the discretionary and contextual expertise to be able to apply that knowledge to a piece of novel research.

Reputational risk

Intellectual property

Loss of revenue

Fairness, Accountability

Breach of confidentiality

The use of a GenAI tool for peer review likely breaches the confidentiality of the author(s) and agreements entered into with the journal/publisher.

Privacy and security

Definitions

The following definitions apply for these guidelines. These are in addition to the definitions outlined in the Research Policy and its procedures.

Bias and transparency refer synthetic persona.

Document use of GenAI includes:

tool name and version
time and date of use
text or screenshots of prompt or query
generated text response
follow-up queries and responses
name of the person who queried
links to original chats (as these can fail, so consider archiving them elsewhere).

Generative AI (GenAI) is defined in Schedule 1, Student Rules.

Hallucination is when a large language model provides incorrect or misleading results, however, appears convincing.

Synthetic persona (synthetic representation) may reinforce biases and particular care must be taken in representation of marginalised and underrepresented groups. Synthetic data should not substitute for engagement with participants or stakeholders. Where synthetic representations are used to summarise or distil key findings or data, there must be an approach to validate these and this use should be transparently reported.

Systematic review requires stringent processes to be followed to control for biases and to ensure transparency. Although GenAI tools can be used for certain tasks within this process (for example, semantic analysis of key texts to help identify terms for search strings and automating extraction of data from sources), current international guidelines and bodies urge caution in this space until a consensus is reached regarding responsible use of AI in systematic reviews. Key concerns include preserving trust, accuracy, validation and integrity in the process. Any GenAI usage must be disclosed and detailed in the methodologies section in accordance with publisher or journal policy. Further information is available at UTS Library: Systematic and other reviews.

Approval information

guidelines contact	Deputy Vice-Chancellor (Research)
Approval authority	Deputy Vice-Chancellor (Research)
Review date	2026
File number	UR25/839
Superseded documents	New guidelines