# Artificial intelligence policies in neurology journals: a cross-sectional analysis

## TL;DR

This study examines AI policies in top neurology journals, finding widespread but inconsistent guidelines for AI use in research and publishing.

## Contribution

The study provides the first comprehensive analysis of AI policies across top neurology journals, revealing gaps in standardization and alignment with ethical frameworks.

## Key findings

- 97% of the top 100 neurology journals have AI-related policies, but these policies are inconsistent.
- Few journals reference established ethical frameworks like ICMJE or COPE, and only one references AI-specific guidelines like CONSORT-AI.
- AI-assisted writing is widely permitted, but permissions for AI-generated content and images vary significantly.

## Abstract

Artificial intelligence (AI) is increasingly used in neurology research and scientific publishing. However, concerns regarding authorship, transparency, and ethical oversight have prompted journals to establish policies governing AI use. The objective of this study was to characterize the presence and content of AI-related author-guideline policies across the top 100 neurology journals and to evaluate their alignment with established editorial frameworks and AI-specific reporting guidelines.

We conducted a cross-sectional analysis of the top 100 neurology journals. Data were extracted from each journal’s Instructions for Authors and included policies regarding AI use, disclosure requirements, authorship restrictions, and permissions for AI-assisted writing, AI-generated content, and AI-generated images. References to ethical frameworks and AI-specific reporting guidelines were also recorded. Associations between AI policies and journal metrics were assessed.

Of the 100 journals examined, 97 included an AI-related policy. Nearly all journals prohibited AI authorship (97%) and required disclosure of AI use (96%). AI-assisted writing was widely permitted (93%), whereas permissions for AI-generated content (77%) and AI-generated images (37%) were more variable. Endorsement of AI-specific reporting guidelines was rare, with only one journal referencing CONSORT-AI or SPIRIT-AI. Few journals cited established ethical frameworks, including the International Committee of Medical Journal Editors (ICMJE; 14%), the Committee on Publication Ethics (COPE; 26%), and the World Association of Medical Editors (WAME; 10%). No significant correlations were identified between AI-related policies and journal metrics.

AI-related policies are common among neurology journals but remain heterogeneous and inconsistently aligned with established ethical and methodological standards. These findings highlight opportunities to strengthen transparency and research integrity as AI becomes increasingly integrated into neurological science. Neurology journals should consider adopting standardized requirements for AI-use disclosure and explicitly endorsing AI-specific reporting frameworks to harmonize expectations and improve reproducibility.