You're working on a benchmark that's supposed to allow, I guess, us to have a better understanding of how well artificial intelligence understands these interfaith relationships, right? So in the same way that these folks are getting together with their peers and having these conversations, you're looking to take that same information and see how well these AI models understand these same types of religious questions.

Yeah.

The assumption is that people who are both interfaith curious and maybe interfaith skeptical are going to take their questions to LLMs, to this AI space, regularly.

And there really isn't right now any evaluation for how those questions are being answered.

And so we were... supported from ACLS, the American Council of Learned Societies, to have a doctoral student join us as a summer intern to help me really flush this project out.

And so I sort of had a hunch about this.

And so the first thing I asked Jake, our intern, to do was to sort of do a survey of the literature and just sort of see what's out there on faith and AI and make sure that my instinct and my kind of searching around this was right.

Is there really a way to erase the bias that we just have as a people that becomes part of the underlying things that the artificial intelligence itself is using?