Can Machines Learn to Worship? Artificial Intelligence in Practical Theology

Will artificial intelligence generate its own religion? Will machines be able to pray and to preaching? And if so, what kind of religion will evolve and how does worship and preaching by machines look like? For Machines to learn to worship, we may have to train machines to learn the structures and contents of worship. In the project *computational religious rhetoric and ritual* we apply machine learning techniques to the study of liturgy and preaching. In our paper we will address the feasiblity of the project along three lines of reasoning. (1) the methodical approach of employing deep learning strategies to classify worship services (supervised methods) and to generate speech-to-text transcriptions (unsupervised or self-supervised methods); the methodological opportunities regarding the type(s) of research questions that can be studied with help of big data analyses; and the theological issue concerning the nature of preaching and worship when technology gains productive agency in Christian worship.