Guiding Responsible AI Use in Scholarly Publishing and Communication: Introducing the L-O-C-A-D Checklist

As artificial intelligence (AI) and large language models (LLMs) become embedded in research and publishing workflows, the scholarly community faces growing pressure to ensure their responsible and transparent use. The L-O‑C‑A‑D (Limitations, Ownership, Confidentiality, Accuracy, and Disclosure) Checklist, developed at EDP Sciences, offers authors, reviewers, and editors a practical, evidence‑based framework to assess and document their use of AI tools throughout the research and publication process.   

The checklist builds on the conceptual model proposed by Dr. Zhou in his Scholarly Kitchen article on the risks, rewards, and responsibilities of AI in peer review, broadening it into an operational, user-friendly tool suitable for diverse research contexts. The L-O‑C‑A‑D Checklist integrates guidance from the European Data Protection Board (EDPB), the Committee on Publication Ethics (COPE), the European Association of Science Editors (EASE), and the STM Integrity Hub to create a consolidated, practice‑oriented resource.  

The L-O‑C‑A‑D Checklist helps users identify and mitigate key risks associated with AI systems, including training‑data opacity, model drift, hallucinations, privacy vulnerabilities, copyright ambiguities, and accountability challenges. It prompts users to assess data quality, avoid sharing confidential or unpublished material, verifying model versions, and ensure meaningful mandatory human oversight. It provides structured guidance for transparent AI disclosure.  

By promoting transparency, reproducibility, and human accountability, the L‑O‑C‑A‑D Checklist supports the principles of Open Science and strengthens integrity within AI‑augmented scholarly communication. This poster introduces the checklist, its rationale, and its practical applications for responsible innovation across the research ecosystem.

This checklist is intended as a practical resource for the scholarly community and to evolve through collaborative feedback. Comments and suggestions are welcome at selina.labarbera@edpsciences.org