large language models on religious texts ... flooding the internet, claiming to offer instanteous spiritual insights. What could go wrong?
The author quotes religious leaders concerned that the work of study and reflection is undermined; his own main worry seems to be that users might be misled by the "hallucinations" of all generative AI. But he doesn't ask whether these models are mining the right data. That they might not be is revealed in a telling quotation from the director of a non-profit called AI and Faith.
Nothing new in theology since Thomas Aquinas? I posted this on the Facebook page of our religious studies alums, and one was good enough to put into words exactly how problematic this is.
Theologians and other theoreticians of the world’s religions will be very surprised to hear Mr. Graves’ news that their religions of study all froze in time 800 years ago. The fact that he thinks new theology must come from AI—as opposed to, say, all the people who have been kept out of the academy forever and are now clawing their way in—or, say, the lived experiences of the billions of religious practitioners—is so outrageous that it’s kind of blowing my mind. I can’t imagine making such a bold claim about something I know nothing about.
Thank you, Helena!
