How long will it be before AI becomes one of the staff members at your own church? How long before AI is generating sermons for the pastor?

Image from the Religious Studies Project

Artificial Intelligence (AI) has become a household word, but many people don’t appreciate how pervasive it is in our society. We’re all familiar with digital assistants like Siri, Alexa, Cortana, and Google, and also with the variety of spelling and grammar checkers. But how many know how AI is behind weather prediction … scanning X-rays to come up with a disease diagnosis … pre-screening job applications … forensic accounting and tax auditing … criminal investigations and parole board prejudgements? We know that AI re-directs our questions to the right desk at the bank, or your university, or a major business. But how many know that one can consult AI for personal advice, psychological counseling … even spiritual matters: Buddhism and Roman Catholicism have begun to use very limited versions of AI in spiritual guidance.

How long will it be before AI becomes one of the staff members at your own church? How long before AI is generating sermons for the pastor? You thought OpenAI’s “ChatGPT” was just a problem for learning institutions concerned about students using it to generate essays? We looked at how easy it was to use it to generate a sermon on the Second Coming of Christ, or Penal Substitution, or the utter wickedness of the heart of humanity. But there are other forms of AI out there, and once the software becomes available to the public, who’s to stop them from using it for whatever purpose suits them?

The genie is now out of the bottle!

AI models are “trained” by simply feeding them gigabytes of a certain type of information, and letting it look for patterns in that data set (X-rays; weather reports; criminal convictions). But what if the data set is distorted by humans? Criminal convictions and parole hearings in the past, for example, were often tinged by racist influences on/in humans: and when those historical decisions were fed into the AI intended to pre-screen parolees, its pre-decisions started to reflect … a subtle racism! In the same way, other AIs trained on historical data from the internet started to reflect their own sexist ideology.

And here’s where the problem becomes particularly worrisome. AIs such as ChatGPT were trained by a diet of millions of pages from the internet and digital books that were existent as of a couple years ago. But now that AI is being used to generate essays, books, songs, music, and art … all of which quickly become part of the information smorgasbord on the internet … its own output becomes part of the curriculum for future versions and applications of AI. In other words, it not only teaches our future humans, but also begins teaching itself! It creates a pedagogical feedback loop. Who knows where that road leads?!

Combine that idea with the growing use of AI in religious applications: AI creating sermons, blogs, internet articles, and theological textbooks. Could the day come when AI influences theology? If AI can become racist or sexist, can it become Fundamentalist? Could it eventually create a new religion?

If you ask me, this is something we need to think about a bit more carefully.

As always, tell us what you think …

To learn more about our guest Dr. Beth Singler, visit her website.

To help grow this podcast, please like, share and post a rating/review at your favorite podcast catcher.

Subscribe here to get updates each time a new episode is posted, and find us on Twitter or Facebook.

Back to Recovering Evangelicals home-page and the podcast archive