From keeping abreast with the perception of artificial intelligence (AI) for several years now, I’m aware that the general population remains pessimistic about AI applications in certain components of society. One such area is politics, where controversial ideas such as AI voting systems have garnered prominent criticism 1. While on the phone with my mother a couple of nights ago (thanks mom!), I had the idea of mixing AI with a similarly contentious area: religion. Specifically, mixing AI and Christianity.
My original idea, thought of during this very phone call, was to develop a recommender system for bible verses. I’ve seen the idea of the bible being a ‘living book’ perpetuated since I was little, and so the idea of ‘assisting’ it through a recommender system seemed almost blasphemous at first. Oddly enough, my mom thought it was a good idea, mentioning how it could help readers find relevant passages faster.
My interest was quite piqued by this idea of using AI to bolster religion, so I looked around to see what, if anything, had been done in this realm so far. A cursory search on Google Scholar yielded some results.
I first found an extremely similar presentation of my original idea, a recommender system for bible verses. A Linguistics Master’s Thesis out the University of Washington in 2018, the thesis presents an approach for recommending verses from the King James Bible, given a paragraph query as input 2.
Also in the CS field, a 2018 paper out of Indonesia presents the idea of developing a web application to help Christians study and understand the Bible 3. It is clear that this idea has sprouted in a small group of minds around the world, but no public-facing product has been developed yet.
More in the humanities realm, there are a selection of journal articles published about how AI might interact with religion. Apocalyptic AI: Religion and the Promise of Artificial Intelligence ,out of the 2008 edition of the Journal of the American Academy of Religion, asserts a major merge between perspectives of “end-times” conditions between scientific and religious figures 4. Geraci, the author, defines Christian apocalyptism as having three primary components: alienation within the world, desire for the establishment of a heavenly new world, and the transformation of human beings so that they may live in that world in purified bodies (Geraci 2008). He believes AI can provide for these three tenets by providing a virtual world inhabited by intelligent machines and human beings who have left their bodies. In true science-fiction style, he believes AI will eventually allow humans to download their consciousnesses into machines, possess enriched mental abilities, and through their infinite replicability, immortality.
Another, more recent 2021 paper in Religions called AI in Religion, AI for Religion, AI and Religion: Towards a Theory of Religious Studies and Artificial Intelligence was more aligned with the questions I’m posing in this article 5. The author here, Randal Reed, asserts that AI will allow researchers to understand religious practices, beliefs, and texts in new ways and firmly believes AI-based applications will be produced in the near future. Reed establishes two types of research that could be assisted by AI: (1) insights into the object of religious beliefs, texts, or practices (through classification and prediction), and (2) studying the thinking of AI itself.
Overall, it is clear that people have recognized how AI might impact religion in the future, but actual developments are just starting to mature (in 2019, with the bible verse recommender system, it seems). I am interested in pursuing this topic further, and working towards further NLP-based projects in this realm.
Finally, I present some more outlandish ideas at this intersection (in other words, the most interesting!).
- Generating new bible verses with a cutting-edge natural language model like OpenAI’s GPT3 (Generative Pre-trained Transformer 3).
- Creating a context-shifted bible — i.e. porting stories and parables into the context of modern American society.
- Producing an averaged version of the Bible, from 30 or so of the leading translations. I think this would be an interesting project because I firmly believe that every translation is an interpretation.
Overall, I think this is a really underexplored area of research with some super cool societal implications. I think a well-written paper on this topic would likely provoke some backlash, but, what impactful research doesn’t?
Footnotes
-
https://www.centreforpublicimpact.org/insights/good-bad-ugly-uses-machine-learning-election-campaigns ↩
-
J. A. Mathias, “Scripture Recommendation Using Document Structure,” p. 46. ↩
-
A. S. Kuswandi, I. N. Sandjaja, and A. Setiawan, “Pembuatan Aplikasi Electronic Guided Bible Study Process dengan Biblical Recommendation System,” Jurnal Infra, vol. 6, no. 1, Art. no. 1, Jan. 2018. ↩
-
R. M. Geraci, “Apocalyptic AI: Religion and the Promise of Artificial Intelligence,” Journal of the American Academy of Religion, vol. 76, no. 1, pp. 138–166, Feb. 2008, doi: 10.1093/jaarel/lfm101. ↩
-
R. Reed, “A.I. in Religion, A.I. for Religion, A.I. and Religion: Towards a Theory of Religious Studies and Artificial Intelligence,” Religions, vol. 12, no. 6, p. 401, May 2021, doi: 10.3390/rel12060401. ↩