Presentation Information
[U03-03]Generative AI in Scholarly Publishing: A Mission-Driven Framework for Responsible Innovation★Invited Papers
*Kristen Averyt1, Matt Giampoala1, Mia Ricci1 (1.American Geophysical Union)
Keywords:
generative AI,publishing,ethics
The rising use of generative artificial intelligence (GenAI) tools by the research community for publication-related tasks introduces new risks and exacerbates perennial issues including bias, hallucinations, plagiarism, and more. For AGU Publications, our strategy to navigate these challenges and effectively leverage GenAI is underpinned by a framework based on five organizational and ethical principles. Our approach to integrating GenAI into AGU publications is community-centered and embeds AGU's core values in how we deploy the technology. Here, we present our philosophical framework for evaluating Large Language Model (LLM) training options and highlight our collaboration with the European Space Agency’s EVE Project. We also outline other initiatives, including AGU’s proprietary Reviewer Finder Tool and partnerships to expand manuscript reviewer pools and to help identify problematic submissions. By integrating new research integrity tools and establishing rigorous GenAI policies for authorship and review, AGU remains committed to upholding the highest quality standards while exploring the frontier of AI-enabled innovation.