Walking the Middle Path : Problem, Policy, and Adaptation to Generative-AI in Academic Publishing.

Abstract

The arrival of publicly-accessible generative-AI tools, such as OpenAI's ChatGPT, has significantly disrupted standard practice in education, research, and publishing. Academic publishers and journals, in particular, have responded by adopting policies which reflect both how they perceive the problems that generative-AI tools pose, and what kinds of solutions address these issues while remaining consistent with key scholarly values: integrity, transparency, and the primacy of human authorship. This article surveys the current state of generative-AI in academic publishing-including relevant policies at the publisher and journal-level-in anticipation of the JBI's forthcoming policies concerning generative-AI disclosure in manuscript submission and peer-review. Mirroring the emerging policy consensus concerning generative-AI use in academic publishing, this paper proposes that 'AI-disclosure statements' constitute an adaptive 'middle path' between prohibition and unrestricted permissiveness. Beyond this, though, we argue that AI-disclosure statements are valuable for their potential to codify desirable norms governing generative-AI use in publishing, and so to subtly influence the underlying 'academic form of life' for the better. We conclude by discussing one possible outcome of this influence, suggesting that it may represent a solution to problems raised by recent criticism of AI-disclosure statements.