I’ve been thinking about a design problem that seems especially difficult for religion and spirituality AI, such as YeshiAI and similar tools:
How personalized should an AI become when it is helping someone with reflection, prayer, journaling, study, meditation, or spiritual practice?
The more personalized it becomes, the more useful it may feel. But it also starts influencing what the user focuses on, which questions they ask, and which practices they repeat.
At that point, is it still just a tool?
There seems to be an important difference between AI helping someone:
- reflect on a teaching
- organize questions
- journal
- stay consistent with a practice
and AI:
- telling someone what they should believe
- presenting one interpretation as correct
- replacing clergy, scholars, teachers, or community
The difficult part is that even “neutral” personalization can shape someone over time.
So where should the boundary be for spiritual and religion-focused AI systems such as YeshiAI?
Can they responsibly become highly personalized around someone’s religious or spiritual path while still avoiding authority?
Or is personalization itself already a form of influence that needs stronger limits?
I’d be interested in how people here would design guardrails for this kind of system.
[link] [comments]