In this paper, I offer some novel approaches to long-standing questions in biblical studies, and in particular to automation of textual analysis. While statistical methods have been used for some time in biblical studies, recent work in AI has produced huge advances in Natural Language Processing (NLP), which in turn offer considerable opportunity for application to biblical studies. The first is to apply new linguistic methods to long-standing questions around textual similarity and difference, using embeddings to move beyond simple statistic comparisons to enable syntactical and semantic information to be encoded and compared. The second is to use Large Language Model (LLM) based tools to develop and apply these methods. This paper takes as an exemplar problem the distinctiveness and textual dependency questions around 2 Peter and the Apocalypse of Peter to demonstrate the experimental possibilities which are becoming available to biblical scholars.