Acknowledging and referencing the use of AI tools in research writing

The acknowledgement and referencing of generative Artificial Intelligence language models in research and academic writing is a dynamic field, lacking a consensus on best practices. Numerous educational and research publications have shared studies and articles discussing AI ethics and its implications in scientific writing, underscoring the necessity for transparency and ethical consideration. These publications often underscore the absence of clear guidelines and best practices in this area. However, given that scientific and academic content is primarily educational and informative, the values of accuracy, transparency, and authenticity are paramount in this field for the information to be deemed trustworthy.

Therefore, when employing AI to generate informative content such as blog posts, news articles, or specific purpose material like marketing or advertising, it is crucial to clearly disclose this and provide a disclaimer. This is because AI-generated informative content may not always be accurate, factual, or comprehensive. To prevent any misinterpretation and as a matter of transparency, it is essential to disclose that the content was generated by AI.

For most organisations, this has necessitated the revision of their authorship and editorial processes and policies to accommodate the use of AI tools in generating their publications. Some organisations and academic institutions have outrightly prohibited the use of AI in any form of personal or presentational writing. Conversely, other organisations permit limited use that adheres to specific guidelines established by the institutions. A complete ban on the use of generating or editing publications might not be enforceable as it is possible to alter AI-generated text to mask it. Moreover, even with AI-text detecting tools, these still provide unreliable methods of detecting AI content. A viable solution at this stage would be to permit the use of AI-generated publications but under stringent principles and guidelines to ensure the transparency and integrity of any publication.

Some minimum requirements currently being used for educative writing and informative publications include:

• Provide the name and version of the generative AI system used and the company that made the system.
• Indicate a brief description of the context in which the tool was used and the relevant prompts used to generate a response.
• The date your output was generated and how you changed it for use or incorporated it into the main publication.
• A link to the chat generated or a compilation of all the generated responses as appendices.

Most organisations nowadays add a small explanation on the disclaimer pages of their publications on the use of AI-generated content within the content. Others have allowed disclosure of AI use in the introduction or methods section of their publications with detailed descriptions of prompts used and indicating the areas of the text affected. Whether disclosed in in-text citations, references or appendices, it is essential to be clear on your AI usage and make it easy for others to read. Furthermore, let your disclaimer or citation remain consistent with the style and tone of your publication.