Historical maps are rich cultural artifacts that pose significant challenges for computational analysis. Maps contain a complex assortment of textual and visual elements, including diverse annotations, nonstandard layouts, and numerous symbolic representations. These features complicate the application of conventional optical character recognition (OCR) technologies. This presentation explores the feasibility of using contemporary artificial intelligence (AI) tools to identify handwritten placenames and topographic symbols—such as churches and fortifications—in early modern cartographic materials. The project focuses on the unprinted works of Danish cartographer Johannes Mejer (1606–1674), whose maps can offer unique insight into 17th-century Nordic geography and culture; for instance, Mejer was among the first to chart the region of Skåne prior to its acquisition by Sweden. The corpus comprises over 200 digitized items from the Danish Royal Library, including maps, sketches, and handwritten documents. These copious materials has not been systematically transcribed or analysed thus far, presenting an opportunity to apply AI-based methods to unlock their content. To evaluate the feasibility of currently available AI technologies for handwritten text recognition (HTR) and symbolic image recognition with respect to such historic maps, the project experiments with several AI-drive tools. Initial experiments have been conducted using Transkribus (German Giant model), HTRFlow, and several large language models (LLMs), including ChatGPT, Claude, and Gemini. Transkribus has shown limited success in accurately identifying handwritten placenames embedded within cartographic contexts. LLMs are being evaluated for their potential – already demonstrating impressive results; a testing environment is now under development to facilitate the systematic comparison of performance by different LLMs as well as integration with Transkribus and ground truth. DHNB 2026 — Book of Abstracts 166 This work-in-progress explores feasibility for applying of AI towards complex historical documents. We invite feedback and collaboration from researchers with experience in machine learning, historical cartography, and archival digitization.

 Large Language Models (LLM s) are increasingly capable of generating Islamic religious discourse perceived as meaningful by human recipients. This article examines one dimension of this capacity: the ability of LLM s to perform istishhād – the citation of Qurʾanic verses to support theological positions – when role-playing six Muslim personas representing distinct positions in contemporary Islamic discourse. Drawing on 480 LLM-generated responses from four models (GPT-4o, GPT-5, Claude 4 Sonnet, and Gemini 2.5 Flash), accessed via API and web interfaces, the article analyses 2,941 Qurʾanic references, examining citation frequency, verse distribution, and cross-persona similarity. Results show that LLM s produce theologically differentiated citation patterns broadly consistent with known positions in Islamic discourse. A secondary finding concerns substantial variation in citation density across models and access channels, raising methodological questions about reproducibility. The article concludes that non-human Islam, while mechanistically novel, exhibits structure, patterns, and theological coherence warranting sustained academic attention.

This article presents a proof-of-concept study on the use of generative AI-based chatbots as synthetic dialogue partners in non-confessional religious education (RE) in Swedish public schools. Addressing the challenge of representing religious diversity without reinforcing stereotypes, the study introduces IRIS (Interactive Religious Identity Simulator), a customizable chatbot designed to simulate diverse religious and non-religious personas. Drawing on theories of cognitive learning and mediated intergroup contact, the article outlines the theoretical foundations, technical construction, and exploratory test cases of IRIS. The findings suggest that AI-mediated dialogue has potential to engage pupils in experimental learning that guides them beyond religious stereotypes.

This article provides a methodologically oriented case study of the application of Large Language Models (LLMs) in Islamic studies. Its empirical focus is on 20 thousand e-fatwas from the website askimam.org. The study demonstrates how LLMs, here represented by the open-source tool BERTopic, can significantly augment a process of distant reading. The research focuses on a topic modelling revealing insights into recurring concerns evident from the questions posed by believers in fatwas. It also provides an example of further, post-topic modelling and analysis focusing on thematic variations across questions from different geographical regions. The article argues for the potential of LLMs to reshape research methodologies, offering faster, more comprehensive, and nuanced analyses of large datasets.