Artificial Intelligence
Artificial Intelligence
The introduction and availability of artificial intelligence (AI) technology and large language models (LLM) have created opportunities in scholarly research and the publishing environment, but have also presented challenges with respect to transparency, credibility, and accountability. Communitas recognises the benefits of these tools that relate to efficiencies, but advocates for the responsible use thereof.
Authors should be aware of the limitations and possible biases of these tools. Irrespective of the tool used or the context thereof, authors alone remain fully responsible for the scientific integrity of their submissions, including for any breach in publishing ethics, such as copyright infringements.
Communitas adheres to the ten core principles of UNESCO’s human-rights centred approach to the ethics of AI as set out in UNESCO’s Recommendation on the Ethics of Artificial Intelligence (UNESCO 2023). Communitas prioritises humans as authors. AI technology, large language models and other technologies and software are tools.
This policy guides the use of such tools pertaining to submissions to Communitas. This policy is informed by the guidelines of the Committee on Publication Ethics (COPE).
Authors
All authors are fully responsible for the originality, validity, and integrity of the content of their manuscript. Authors submitting manuscripts to Communitas attest in the Publishing Agreement, inter alia, that the manuscript is their original work and does not contain any unlawful content and does not infringe any existing third-party copyright, moral right or other intellectual property rights; and that the work of others has been appropriately attributed. As such, AI-generated content will not be considered for publication. Any submission found to include AI-generated content will be declined [or retracted if already published]. AI technology and models may not be listed as an author. The use of an AI for any purpose must be clearly declared. Any undeclared use or attempt to hide the use of AI tools will be treated as a breach of publication ethics.
The above mandate notwithstanding, AI and large language models may be used to edit descriptive (non-data) images, revise and edit writing, and seek and summate existing literature. Such use and the extent thereof must be declared at the time of submission in the cover letter and detailed in the Methods.
The declaration of such a use should include the name, version, and manufacturer of the tool used, and the date on which it was accessed, for example: (Chat GTP 3.5, Version 28 August 2023, Open AI, accessed 16 September 2023)
The “prompt” or plain-language instruction used must also be provided, either in the Methods section of the manuscript or as supplementary material to the manuscript.
Where AI tools or large language models have been used, for example, in the conception and design of a study, editing of non-data images, or in editing or revising the writing, such contributions do not meet the criteria for authorship.
The use of tools specifically for spelling and grammar checks, similarity checking, and reference management is fully permitted and does not need to be disclosed.
Reviewers
Reviewers are responsible for the content of their review reports, including adhering to the principle of confidentiality. AI-generated review reports will not be considered. All submitted manuscripts and correspondence with the Editorial Office should be treated as confidential and may not be shared in any way. Reviewers may not upload full manuscripts or any part of a manuscript into a large language model or any generative AI tools.
Reviewers who use an AI tool as a resource for peer review, in a way that does not violate confidentiality, must disclose and detail this use in the review form and provide the name, version, and manufacturer of the tool used as well as the prompt provided.
Editors
A similarity check is performed on submitted manuscripts that undergo peer review. Communitas does not use any tools to replace the work of peer reviewers in the evaluation of manuscripts, and editors of Communitas do not use AI-based tools in the writing of their decision letters.
Below is a summary of these stipulations as applicable to authors:
Example
Writing/generating any part of a manuscript
e.g. “Write 3000 words on [specific topic], covering key concepts, recent developments, methodologies, and potential future directions.”
e.g. “Write an Introduction to the below text and add key references.”
AI may be used: NO
Used must be disclosed: not applicable
Example
Creating images
AI may be used: NO
Use must be disclosed: not applicable
Example
Modifying data images, e.g. gel images, micrographs
AI may be used: NO
Use must be disclosed: not applicable
Example
Conception and design of study
e.g. “Evaluate this study protocol to identify possible confounding factors.”
AI may be used: YES
Use must be disclosed: YES
Example
Editing, translating and summarising papers or large sections of writing
e.g. “Provide recommendations to improve readability of the text.”
e.g. “Edit the text to reduce to 250 words while preserving content, intention and clarity.”
AI may be used: YES
Use must be disclosed: YES
Example
Editing non-data images
AI may be used: YES
Use must be disclosed: YES
Example
Grammar checking and copyediting tools
AI may be used: YES
Use must be disclosed: author’s discretion
Example
Similarity checking tools
AI may be used: YES
Use must be disclosed: author’s discretion
Example
Reference managers
AI may be used: YES
Use must be disclosed: author’s discretion
The use of AI tools in scholarly publishing is expected to evolve, and this policy will be revised as needed.
This policy has been reproduced/adapted from the South African Journal of Science. The South African Journal of Science and its publisher, the Academy of Science of South Africa (ASSAf), accept no responsibility for the accuracy, applicability, or outcomes resulting from the use of these policies by other journals.
