Can a Therapist Use AI? The Prudential Line Every Clinician Must Hold
Father Robert Spitzer's recent commentary in the National Catholic Register establishes that generative AI is not intrinsically evil but demands prudential restraint. For Catholic Christian clinicians, that framing opens a precise question: where does legitimate use end and clinical abdication begin? The answer turns on what therapy actually requires of a human person.
The question behind the question
A therapist reviews a client's case notes at 10 p.m. and wonders whether an AI tool could help organize the material, flag relevant literature, or draft a session outline for the morning. The question feels practical. It is, in fact, ethical.
Father Robert Spitzer, writing in the *National Catholic Register*%2C%20because%20the%20very%20use%20of%20generative%20AI%20does%20not%20always%20directly%20cause%20unjust%20irreparable%20harm%20or%20the%20negation%20of%20human%20dignity.), offers the theological coordinates: "generative AI is not intrinsically evil... because the very use of generative AI does not always directly cause unjust irreparable harm or the negation of human dignity" (Spitzer, 2026). Catholics, he argues, "are free to use generative AI so long as they prudentially exercise caution and restraint" (Spitzer, 2026). That is a broad permission. The harder work is mapping what prudential caution actually means inside a clinical relationship.
Where AI belongs in the consulting room
Some legitimate uses may be administrative and preparatory: scheduling, billing queries, literature searches, and perhaps note-drafting after a session ends. These functions stay outside the clinical relationship itself. A language model that summarizes DSM criteria or retrieves a published protocol is functioning the way a well-organized library does. No client is harmed; no clinical judgment is delegated.
The line that cannot be crossed involves assessment, interpretation of a client's disclosure, or any substitute for clinical supervision. Those acts require moral and relational accountability. They require a person who can be wrong, who can be held responsible, who bears the weight of another's suffering as a fellow human being. The American Psychological Association's own ethics code anticipates this exact boundary: psychologists "retain responsibility for the appropriate application, interpretation, and use of assessment instruments, whether they score and interpret such tests themselves or use automated or other services" (American Psychological Association, 2017). The obligation does not transfer because the intermediary is software.
Spitzer names the risk directly: "AI's mimicry of human interpersonal relationships can even encourage immoral and suicidal behavior" (Spitzer, 2026). A clinician who routes a client's disclosure through an AI tool is not using a library. She is introducing a system that performs empathy without holding any of its consequences.
What a therapist does that AI cannot
Therapy works because a human being bears witness. The therapeutic relationship is not an information-exchange protocol; it is a moral act performed by an embodied person who brings conscience and situated judgment into contact with another person's suffering. As Spitzer points out, "AI will never be human" (Spitzer, 2026). Genuine empathy and moral accountability cannot be reduced to pattern-matching across training data, however vast.
A distinction worth pausing on: unlike a calculator or a hammer, AI is not a neutral tool, because it presents itself as a relational agent; some will even "greet" you in the morning. That distinction matters most when the person on the other side of the screen is already vulnerable. A clinician who outsources clinical reasoning to a system trained on undifferentiated mass data has not gained a resource. She has substituted a simulation for judgment.
Formation before function
The question for any Catholic Christian clinician is not whether AI is evil. It is whether she retains the prudential judgment, relational presence, and moral accountability that make therapy therapeutic. Tools that sharpen preparation belong. Tools that replace the person do not.
Spitzer's framework, applied to clinical practice, points toward the same standard that governs any professional decision: prudence runs contrary to expedient, shooting-from-the-hip reactions, and its fruit is a sound conscience. That conscience is what a client actually needs in the room.