If you have or you plan to use Artificial Intelligence (AI) tools, including Large Language Models (LLMs) like ChatGPT, ensure that you check the relevant journal policies about use and acknowledgement. 

Many scholarly publishers agree on the following general principles:

1. An LLM can not be a study author

"Large Language Models (LLMs), such as ChatGPT, do not currently satisfy our authorship criteria. Notably an attribution of authorship carries with it accountability for the work, which cannot be effectively applied to LLMs."   Nature Authorship Policy

"Authors must be aware that using AI-based tools and technologies for article content generation, e.g. large language models (LLMs), generative AI, and chatbots (e.g. ChatGPT), is not inline with our authorship criteria. All authors are wholly responsible for the originality, validity and integrity of the content of their submissions. Therefore, LLMs and other similar types of tools do not meet the criteria for authorship." Taylor & Francis Editorial Policy on Authorship

2. Content generated with assistance from AI tools must be acknowledged in sections of the paper other than the author list.

"Use of an LLM should be properly documented in the Methods section (and if a Methods section is not available, in a suitable alternative part) of the manuscript."  Nature submission guidelines

"Any assistance from AI tools for content generation (e.g. large language models) and other similar types of technical tools which generate article content, must be clearly acknowledged within the article. it is the responsibility of authors to ensure the validity, originality and integrity of their article content. Authors are expected to use these types of tools responsibly and in accordance with our editorial policies and principles of publishing ethics."  Taylor & Francis Editorial Policy on Authorship

Some journals, including the esteemed Science, require that permission be obtained from the journal's editors for the inclusion of any AI-generated content,

"Text generated from AI, machine learning, or similar algorithmic tools cannot be used in papers published in Science journals, nor can the accompanying figures, images, or graphics be products of such tools, without explicit permission from the editors." Science Journals Editorial Policies

The manner is which use of AI in a study is described in a paper will vary from discipline to discipline. For example, where relevant, Mathematics and Physical Sciences papers should discuss and reference the following as a minimum, to assign attribution and enable readers to re-create the outcome: 

• Source (e.g. OpenAI/Microsoft, Anthropic/Google, NVIDIA etc. )
• Model (e.g. GPT 3)
• Implementation (e.g. davinci-003)
• Fine-tuning*

*Where the user has fine-tuned the 'inbuilt' knowledge of LLMs based on their own libraries of content via APIs or other processes.

The drafting stage gives you the opportunity to check not only the structure and content of your paper, but to make sure key matters like publisher guidelines, bylines and intellectual property have been properly considered and addressed.

• Quality of your paper - The draft paper should be reviewed and proof-read. This includes receiving feedback on the content of your paper (including title and abstract), as well as having the paper checked for consistency and accuracy amongst the main text, references, tables and figures. Consider those around you who are best placed to review your paper – co-authors, colleagues, supervisors, mentors – and the type of feedback they are able to provide, for example a more technical review versus a grammatical check.

• Instructions to authors from publishers - These are guidelines publishers expect you to follow and include instructions on how to submit the article, the format of the paper, word count, referencing style, acknowledging use of AI tools and similar matters. Investigate and follow publisher/journal style and format requirements.

• Byline - It is important all papers include 'The University of Western Australia' as part of the byline. This ensures the publication is captured in the UWA Research Repository. Contact your School's or Centre's administration staff to confirm other specific requirements.  

• Acknowledgements - make sure that all appropriate acknowledgements (financial, infrastructure, use of AI tools or otherwise) have been made and to confirm with other co-authors that all necessary acknowledgments have been included.

• Intellectual Property - Is the publication disclosing valuable intellectual property that is associated with or has arisen out of the research project? If confidential or proprietary information is disclosed in a publication without the consent of the owner of that information, you may be in breach of your research agreement, and you could cause wider issues for that intellectual property, including destroying it’s patentability.

Always refer back to any governing agreement (see tab Before you publish) and make sure those who need to see and approve your draft publication do so, and that you do not include any information you shouldn’t.  Additionally, disclosure of IP by publication may prevent you later protecting that IP with patent and subsequently the commercialisation of the idea.

Questions about intellectual property? Contact the Research Commercialisation team or the UWA Librarian Support team. 

Email your questions to our friendly library staff.

Students

web.uwa.edu.au/askuwa

HDR Students

hdrsupport-lib@uwa.edu.au

UWA Staff

staffsupport-lib@uwa.edu.au

More contact options are available on the Library Contact us page.