AI Tools in Academic Writing: Guidelines for Ethical Use

The American Institute of Mathematical Sciences (AIMS) acknowledges the growing role of Artificial Intelligence (AI) in academic writing and research. AI tools, such as ChatGPT, offer transformative potential by assisting with idea generation, overcoming writer's block, and streamlining editing processes. However, AIMS emphasizes the importance of understanding the limitations of these technologies and using them responsibly to uphold academic and scientific integrity. Human oversight and accountability remain crucial to ensure the accuracy and reliability of published content. These guidelines support authors in effectively integrating AI tools into their work while maintaining the highest standards of scholarly publishing.

It's essential to differentiate between assistive and generative AI tools. Assistive AI tools, such as Grammarly, Curie, and LanguageTool, enhance content authored by the user by providing suggestions, corrections, and improvements. These tools refine and improve clarity in independently created content. Generative AI tools, like ChatGPT and DALL-E, produce original content, including text, images, and translations. Content primarily created by generative AI is considered 'AI-generated,' even with subsequent human modifications. Understanding this distinction is vital for proper disclosure and responsible use.

While the use of assistive AI tools does not require disclosure, AIMS mandates that all content, including AI-assisted content, undergo rigorous human review to ensure quality and authenticity. Authors must disclose any AI-generated content in their submissions, including text, images, or translations. This disclosure allows the editorial team to make informed publishing decisions. A 'Use of AI tools declaration' section is provided for this purpose, requiring authors to specify the tool used and its application in the work.

When using generative AI tools, authors must adhere to specific guidelines to maintain academic integrity. First, disclose the use of AI tools in the submission. Second, carefully verify the accuracy, validity, and appropriateness of AI-generated content, as Large Language Models (LLMs) can produce incorrect or misleading information. Third, meticulously check sources and citations, ensuring proper referencing. Fourth, appropriately cite AI-generated content following established referencing conventions. Fifth, avoid plagiarism and copyright infringement by confirming that the submission contains no plagiarized material. Sixth, be aware of potential biases in AI-generated text and ensure inclusivity and impartiality. Seventh, acknowledge the limitations of LLMs, including potential inaccuracies and knowledge gaps. Eighth, remember that AI tools cannot be recognized as co-authors, and the author remains responsible for the work. Finally, stay updated on the latest developments and ethical challenges related to AI-generated content.

Certain uses of generative AI are strictly prohibited. Authors must not use generative AI to create or modify core research data artificially. Sharing sensitive personal or proprietary information on AI platforms like ChatGPT is also prohibited, as it may expose confidential data or intellectual property. Editors and reviewers must maintain the confidentiality of the peer review process and must not share information about submitted manuscripts or peer review reports in generative AI tools. Reviewers are also prohibited from using AI tools to generate review reports.

AIMS follows the Committee on Publication Ethics (COPE) guidelines regarding the use of AI tools. COPE emphasizes that AI tools cannot be listed as authors of a paper, as they cannot take responsibility for the submitted work, assert conflicts of interest, or manage copyright and license agreements. Authors must disclose the use of any generative AI tools in the writing of a manuscript, the production of images or graphical elements, or the collection and analysis of data. Authors are fully responsible for the content of their manuscript, including any portion produced by an AI tool, and are liable for any breach of publication ethics.

The guidelines provided by AIMS aim to ensure the responsible and ethical use of AI tools in writing and research, preserving the integrity and quality of academic and scientific publications. By adhering to these guidelines, authors can leverage the benefits of AI while maintaining the highest standards of scholarly work. Disclosure, verification, and awareness of limitations are key to responsible AI integration.

For further information on AI in scholarly work, authors can refer to resources such as the World Association of Medical Editors (WAME) recommendations on chat bots and scholarly manuscripts, the Committee on Publication Ethics (COPE)’s position statement on Authorship and AI tools, and the STM Whitepaper on Generative AI in Scholarly Communication. These resources provide additional insights and guidance on navigating the evolving landscape of AI in academic publishing.