arXiv:2605.28843v1 Announce Type: cross 
Abstract: AI is transforming life sciences research at unprecedented speed, accelerating discovery across protein structure prediction, genome modeling, and drug development (Jumper et al., 2021; Mak et al., 2024). Yet this rapid advancement, coupled with the open science movement, introduces significant dual-use research concerns that have received limited empirical scrutiny. Here we present the first systematic analysis of dual-use research of concern (DURC) content on open preprint servers. We screened ~52,000 bioRxiv preprints (2024-2025) using a hybrid pipeline of lexical filtering and large language model (LLM) evaluation, scoring metadata across nine DURC, three PEPP, and five governance categories aligned with U.S. and Australia Group oversight frameworks. Our analysis reveals that dual-use-adjacent knowledge is routinely present in openly accessible titles and abstracts, often exceeding established risk thresholds even in studies with legitimate public health objectives. While this mapping captures surface-level information diffusion, it does not measure operational capability, downstream misuse potential, or the substantial technical and biosafety barriers that constrain harmful application. We argue that institutional review processes, funding requirements, and preprint platform policies must evolve to incorporate proactive, metadata-level monitoring without compromising scientific transparency. Ultimately, harmonizing controlled-access mechanisms for high-risk methodologies with open summaries of scientific contributions offers a pragmatic framework for governing AI-accelerated biology at scale.

تم إجراء تحليل منهجي لمحتوى البحث ذي الاستخدام المزدوج (DURC) على خوادم ما قبل الطباعة المفتوحة، مما يكشف أن المعرفة القريبة من الاستخدام المزدوج موجودة بشكل متكرر في ما قبل الطباعة على bioRxiv، وغالبًا ما تتجاوز الحدود المعمول بها للمخاطر. قامت هذه الدراسة بفحص حوالي 52,000 ما قبل الطباعة باستخدام خط أنابيب هجين من التصفية اللفظية وتقييم نموذج اللغة.

Se ha realizado un análisis sistemático del contenido de investigación de doble uso (DURC) en servidores de preprints abiertos, revelando que el conocimiento adyacente al doble uso está presente con frecuencia en los preprints de bioRxiv, a menudo superando los umbrales de riesgo establecidos. Este estudio examinó aproximadamente 52,000 preprints utilizando un pipeline híbrido de filtrado léxico y evaluación por modelo de lenguaje.

Une analyse systématique du contenu de recherche à double usage (DURC) sur les serveurs de préimpression ouverts a été réalisée, révélant que des connaissances adjacentes au double usage sont fréquemment présentes dans les préimpressions de bioRxiv, dépassant souvent les seuils de risque établis. Cette étude a examiné environ 52 000 préimpressions à l'aide d'un pipeline hybride de filtrage lexical et d'évaluation par modèle de langage.

A systematic analysis of dual-use research of concern (DURC) content on open preprint servers has been conducted, revealing that dual-use-adjacent knowledge is frequently present in bioRxiv preprints, often exceeding established risk thresholds. This study screened approximately 52,000 preprints using a hybrid pipeline of lexical filtering and large language model evaluation.