Policy on AI Technologies

Image on this page was generated from a text description using Adobe Firefly.

If authors use AI tools in the preparation of their manuscripts, they must ensure that any substantive use of these tools meets high standards of reliability and replicability. This is crucial for maintaining the integrity and credibility of the research. Therefore, when utilizing AI for substantive tasks, authors should provide detailed documentation of the AI tools used, including their versions, settings, and the specific applications. The use of AI can range from non-substantive to substantive contributions. Non-substantive use includes tasks such as spelling and grammar correction, as well as refining the text’s formulation and formatting. Substantive use, on the other hand, involves more complex tasks like data analysis.

If artificial intelligence (AI) technologies, including large language models (LLMs) such as ChatGPT or other AI tools, have been utilized in the preparation of your manuscript or at any stage during the research process, it is necessary to provide detailed information about their use in the “Materials and Methods” section. This should include:

  • Description of AI Technologies: Clearly specify which AI technologies or tools were employed.
  • Purpose and Application: Describe how the AI technologies were applied in your research. This could include their role in data analysis, text generation, experimental design, or any other relevant aspect of the research process.
  • Transparency in Contribution: Provide a detailed account of how the AI technologies contributed to the research outcomes.
  • Methodological Details: Include information on the specific algorithms, models, or software versions used, and any key settings or parameters if applicable.

Examples of AI technologies or tools include[1]:

  • Natural language processing tools such as Chatbots (ChatGPT, Claude) and Text analysis tools (e.g. Grammarly, Paperbot, Quilbot)
  • Machine learning frameworks such as TensorFlow or PyTorch
  • Data analysis tools such as DataRobot or Tableau
  • Computer vision tools such as OpenCV
  • Speech recognition tools such as Google Speech-to-Text
  • Robotic process automation tools such as Elasticsearch
  • Predictive analytics tools such as IBM SPSS
  • Recommendation engines such as Netflix’s recommendation algorithm
  • AI in business applications such as Marketo, Alteryx, IBM Watson for Health

It is also the duty of the authors to thoroughly check all items for correctness for which an AI tool was used. If an AI technology was used during the proofreading or revising stage to improve the language, the original version should also be saved in case this is requested by the Editors for review. The authors are also responsible for ensuring that when uploading any data or text to an AI tool, that all items relating to copyright, privacy and confidentiality have been considered. It is the authors’ responsibility to ensure that all necessary rights to material used have been obtained in advance, and that the work of others has been properly cited in the text and in the references.

Please note that large Language Models (e.g., ChatGPT) or other artificial intelligence tools cannot be listed as authors as “they cannot be responsible for the accuracy, integrity, and originality of the work[2]”, all of which are required to be considered according to the guidelines set by ICMJE.

This approach ensures that AI-enhanced research adheres to rigorous standards, enhancing trust and transparency in published works.

Image Source: The image on this page was generated from a text description using Adobe Firefly.

This can be downloaded as a pdf here.

[1] Source: https://www.linkedin.com/pulse/what-artificial-intelligence-ai-tools-jatin-uaqkf/

[2] Source: https://www.icmje.org/recommendations/browse/roles-and-responsibilities/defining-the-role-of-authors-and-contributors.html