Kalman asks which religious skills AI users should keep practising

An interview about retained skills

D.Z. Kalman has raised a question for educators using AI: which skills do they want people to keep? In an interview published by the United Methodist Church’s Compass podcast on October 14, the Jewish scholar distinguished asking a system to impersonate a preacher from asking it to check a sermon’s grammar.

Kalman also described the choice facing someone who can translate demanding religious texts when software offers a quicker route. He urged attention to what users ask a model to do, including activities that might be acceptable only in moderation. These were reflections in an interview, not announced rules: he said guidelines being developed by a Shalom Hartman Institute team were unfinished. The conversation provides no measured evidence that a use causes skill loss. Source: umc.org

Earlier educational context

The educational question extends beyond congregations. UNESCO’s guidance on generative AI in education and research, first published in September 2023, puts human capacity and pedagogical design alongside ethical validation. That is a useful distinction: an acceptable output and a worthwhile learning process are different things to assess. Source: unesco.org

Christian application: judgment needs practice

For Christian teachers, Hebrews 5:11–14 supplies a formation question rather than a technology rule. The passage rebukes readers who should have progressed enough to teach but still need elementary instruction. Verse 14 describes people whose practice has trained them to distinguish good from evil. Applying that to AI requires interpretation; the passage does not discuss software. Its challenge is whether a learning arrangement gives people occasions to exercise judgment themselves. Hebrews 5:11-14 (NIV)

A classroom question to try

A Bible-study leader could therefore separate preparing an answer from testing one. Before consulting a chatbot, learners might read the passage, identify its immediate context and explain their own interpretation. Afterwards, they could compare the generated response with the text and name what they would accept, reject or investigate. The teacher would assess the explanation, not simply the polish of the final paragraph.

A useful review question would be: could the learner explain this conclusion if the tool were unavailable? That is a proposed classroom check, not a validated test or a requirement from Kalman. It keeps the decision concrete. A congregation can choose where assistance saves effort while reserving enough practice for people to remain responsible interpreters of what they teach.