In “Is Lying Only Sinful in Islam? Exploring Religious Bias in Multilingual Large Language Models Across Major Religions” (arXiv), Kazi Abrab Hossain (BRAC University) and others “introduce BRAND: Bilingual Religious Accountable Norm Dataset, which focuses on the four main religions of South Asia: Buddhism, Christianity, Hinduism, and Islam, containing over 2,400 entries, and we used three different types of prompts in both English and Bengali. Our results indicate that models perform better in English than in Bengali and consistently display bias toward Islam, even when answering religion-neutral questions. These findings highlight persistent bias in multilingual models when similar questions are asked in different languages.”
A podcast on “Artificial Intelligence (AI) in African and Islamic Studies” explores “how artificial intelligence is transforming research practices: How can AI support historical and anthropological research—and where are its limits? And how might AI change the future of the humanities—opening up new possibilities while also raising pressing questions about language equity, digitisation gaps, and academic responsibility?” The episode features a conversation with Frédérick Madore (ZMO Berlin), who “is developing the Islam West Africa Collection—an open-access digital archive of more than 14,500 documents on Islamic print culture and Muslim life in West Africa.”
