Skip to content

Latest commit

 

History

History
32 lines (18 loc) · 4.13 KB

acknowledge-ai.md

File metadata and controls

32 lines (18 loc) · 4.13 KB

Journals and book publishers seem increasingly open to the idea of authors making use of generative AI in the writing process, although most will not allow generative AI models to be credited as co-authors.

In the spirit of sharing resources, I thought it might be useful to provide a template that authors might use when submitting work that has been developed with generative AI. The example below was used for a book chapter but with some minor editing it could serve for an article as well.

Acknowledgement of the use of generative AI in writing this chapter

I acknowledge that I have used MODEL (VERSION) to support the writing process in the production of this chapter. In this I was guided by the recommendations of the NEJM AI editorial board (Koller, et al., 2023).

At NEJM AI, we have therefore elected to allow the use of LLMs. Our two key conditions are first, that the use of LLMs is appropriately acknowledged by the authors... . Second, we require that the authors be completely accountable for the correctness and originality of the submitted work.

As part of this acknowledgement, I would like to describe how I used MODEL support my writing.

To ensure that the chapter was grounded in a set of my own core ideas and perspectives, I provided the model with LIST RESOURCES written by AUTHORS. These documents outlined the central themes, definitions, and questions that the chapter aimed to address, alongside relevant examples. In addition, I supplied MODEL with a comprehensive set of notes that I had created independently, without the use of generative AI. These notes contained relevant concepts, examples, and insights that I had gathered through my own exploration of the topic.

Using these three sets of materials as a framework, MODEL helped me to develop an outline for the chapter. The AI analysed the documents and notes, identifying key concepts and organising them into a coherent structure. This outline served as a roadmap for the subsequent writing process, where I engaged in an interactive dialogue with MODEL, using a series of prompts to delve deeper into specific aspects of the topic. Through this conversational process, MODEL helped me to refine and expand my own thinking. These prompts included, but were not limited to, the following:

I include these prompts as a set of examples...obviously every author will have a unique set of prompts they used in their interactions with the language model.

  • I have attached 3 files that each include a version of a book chapter on TOPIC. The book itself is DESCRIPTION OF PIECE, AIM, AUDIENCE, ETC. I want you to analyse each of the attachments and take the high level ideas as a framework for a draft of the chapter.
  • Given that our working definition of TOPIC for this book chapter includes DETAILS, is reasonable to say that QUESTION?
  • I recognise that an AI cannot XXX in the context that we're talking about here. However, I want to explore the question of what it might look like if an AI could XXX. Playing devils advocate, what would be so wrong with an AI taking over the role of XXX, especially if it turned out that an AI could be more XXX than a human?
  • OK, let's say that the future of XXX in health and social care may include an AI element, that supports and enhances what humans are capable of on their own. Using the attached documents as guidance, let's talk about what this might look like.

Throughout the dialogue, I integrated MODEL's suggestions and ideas into my own writing, serving as a thought partner, and helping me to explore the topic more deeply and consider perspectives that I might have otherwise overlooked. It is therefore not unreasonable to suggest that MODEL's influence is reflected throughout the chapter. However, I am solely responsible and accountable for the information submitted.

Additional resources

  • Koller, D., Beam, A., Manrai, A., Ashley, E., Liu, X., Gichoya, J., Holmes, C., Zou, J., Dagan, N., Wong, T. Y., Blumenthal, D., Kohane, I., & the editors and editorial board of NEJM AI. (2023). Why We Support and Encourage the Use of Large Language Models in NEJM AI Submissions. NEJM AI, 1(1), AIe2300128. https://doi.org/10.1056/AIe2300128