Spiritual-care commentary asks whether AI reassurance amounts to relationship

A question about the encounter

Can a chatbot’s reassuring words become a relationship of spiritual care? A commentary published November 9 in the Journal of Medical Internet Research asks that question through Martin Buber’s philosophy. Authors Fransisca Christiane Warmenhoven, Bauk Wilhelmina Jan Aartsen and Roos Notermans work in Maastricht University Medical Centre’s spiritual-care department. Source: jmir.org

The journal labels the article non-peer-reviewed. Its first preprint appeared October 1. The authors describe a December 2025 Dutch-language exercise involving four AI systems and a scenario about someone keeping vigil beside a dying father. They judged the replies useful for information, reassurance and acknowledging emotions. This was a vignette exercise, not a clinical trial, and the commentary reports no patient outcomes. Source: jmir.org

Their deeper concern is whether such an exchange can provide the mutual presence associated with Buber’s “I–Thou” relationship. They argue that AI lacks the lived experience of mortality that people bring to spiritual care. Yet they end with an open question about whether an I–Thou relationship with AI is possible, rather than claiming to have settled it experimentally. Source: jmir.org

Expert judgments are another kind of evidence

The commentary responds to a July 27 Delphi study by Fabian Winiger and colleagues. That study gathered expert judgments over two rounds, with 83 participants completing the second. Respondents were more supportive of AI assistance with administrative and informational work than with direct relational care; loss of human connection was the highest-ranked risk. Source: jmir.org

Those are professional assessments, not measured effects on patients. The study did not include patient perspectives, and its panel was concentrated in North America and Europe. It calls for research on outcomes in particular tasks and settings. Source: jmir.org

What a ministry should ask next

For a church or chaplaincy team considering a spiritual-support tool, the useful distinction is between the quality of a reply, the recipient’s experience and the continuing responsibility to care. A demonstration can show thoughtful wording. To understand whether someone felt heard requires asking that person. To establish what happens after a difficult exchange requires an actual service arrangement.

Our editorial suggestion is to ask these questions separately: What has been checked in the responses? Whose experiences have been studied? Who is available when the conversation requires a visit, clarification or follow-up? Keep unanswered questions visible rather than treating a persuasive demonstration as evidence for all three. This is a way to examine claims, not a recommendation to introduce a chatbot into patient care.

For Christian readers, Romans 12:9–21 places sharing grief alongside hospitality, generosity, humility and peace. Paul’s instructions concern a community’s conduct, not a test of machine consciousness. Applied to ministry, they ask more of a congregation than producing appropriate words: its members remain responsible for responding to a person’s circumstances. That responsibility should be part of any discussion about a tool’s proposed role. Romans 12:9-21 (NIV)