AI Policy

To ensure the integrity and quality of Hyphae, we require authors, peer reviewers, and staff to abide by our AI Policy explained here. This is not only to protect the material we publish. It is also to make sure everyone’s experience of writing and publishing peer-reviewed articles is just, respecting everyone’s precious time and effort.

Overall principles: When an author publishes work in Hyphae, we must claim copyright of that material to the author/s. If there is AI-generated content in the article, that is non-copyrightable material. Thus, we would be claiming copyright on non-copyrightable material. We cannot do this. So, we must make sure manuscripts have no AI-generated writing (including formulating arguments). In short, the “pipeline of work” must all be done by the person. The manuscript must be an expression of the author/s all along the way, from ideas, to design of the writing, to the writing itself.

What is acceptable:

  • Brainstorming, literature review, outlining, transcribing text from video/audio, translation of languages
  • Polishing your writing, e.g. as an advanced spell check but not editorial work that can shape and change the argument
  • Formatting of your own text
  • Microsoft Excel functions that happen to use machine learning
  • “Stress-testing” arguments or reasoning
  • Exploring research questions – this could be done before the writing phase. But, please be careful that your interaction with AI can shape your ways of thinking iand leave you with biases. We recommend discussing your ideas with peers. At Hyphae we offer Hyphae Conversations and Hyphae Works in Progress precisely for having opportunities to think together in communities of inquiry.

What is NOT acceptable:

  • Having the AI structure one’s argument
  • AI use is “opaque” (gives results that nobody can verify)
  • Author, reviewer, or editor responsibility is delegated to AI
  • This makes the AI perform the role that the person was supposed to do. This wastes everyone else’s time and effort who act in good faith as if the person did the work.
  • Confidentiality/intellectual property is breached

Helpful examples of what is NOT allowed:

  • AI-generated hypotheses, conclusions, or analysis/reasoning for what is supposed to be your article
  • Fabricating data, citations, or results
  • Creating “deepfakes” (that is, photorealistic images)