top of page

The Law Caught Up With the Chatbots. Our Theology Got There First.

Seven states now restrict or prohibit AI from delivering therapy. Illinois and Nevada moved first in 2025. Colorado, Maine, Rhode Island, Tennessee, and Vermont followed in 2026. Utah took a softer path and requires disclosure instead. The statutes differ in detail, but the legislative logic is consistent. Therapy is provided by licensed persons, and software is not a person.


It took a wave of documented harms for secular legislatures to arrive at a conclusion Christian anthropology has held from the beginning. Care for the soul requires a soul.


What the laws actually say


The common statutory pattern is narrow. AI may handle administrative work: scheduling, billing, documentation support, referrals. AI may not make independent therapeutic decisions, generate treatment plans without clinician review, or present itself as a licensed professional. Tennessee prohibits AI systems from representing themselves as qualified mental health providers. Rhode Island bars licensed clinicians from letting AI make independent treatment decisions. Maine classifies AI-delivered therapy as an unfair trade practice.


Read your own statute. Licensure is state law, and this area is moving fast. More than thirty states introduced chatbot bills in early 2026. If you see clients across state lines through telehealth, you may answer to more than one framework.


The theological ground under the legal fence


The Word became flesh and dwelt among us (John 1:14). God's response to human suffering was embodied presence. That is the pattern for every ministry of care that follows.


Counseling, understood Christianly, is one image bearer sitting with another. Paul writes that God comforts us in our affliction so that we can comfort others with the comfort we ourselves have received (2 Corinthians 1:3-4). The chain matters. Comfort passes through people who have suffered and been comforted. A language model has never been afflicted and has never received comfort. It can generate comforting sentences all day while standing entirely outside the economy of grace Paul describes.


Burden bearing works the same way. Galatians 6:2 assumes a bearer with something at stake, someone whose evening is heavier because they carried your grief. Software carries nothing. The cost is part of the care.


There is a fallenness dimension too. Our clients sit across from a counselor who is also fallen, also under grace, also in need of the same mercy. That shared condition carries real clinical and spiritual weight. A chatbot has no sin to confess and no grace to receive. Whatever it offers, it offers from outside the moral universe our clients actually inhabit.


Where AI belongs in a Christian practice


None of this makes the technology useless. In my own work AI earns its keep on the administrative side of the wall: drafting documentation for my review, summarizing literature, producing psychoeducation handouts I then edit, streamlining scheduling and billing. The new statutes permit this and prudence commends it, under two conditions. Everything AI touches gets clinician review before it reaches a client or a chart. And anything involving protected health information runs through HIPAA-compliant tools with a business associate agreement in place.


Informed consent should say what you do. If AI assists your documentation, disclose it. Some states now require disclosure by statute. All of them reward it in a board hearing.


Where it does not belong


  • Diagnosis or treatment planning without your review and signature

  • Crisis response of any kind

  • The session itself, in person or by proxy

  • Anything a reasonable client would mistake for your clinical judgment


As of its November 2025 advisory committee meeting, the FDA had authorized no generative AI device for any clinical mental health purpose. The committee spent that meeting cataloguing the reasons: automation bias, missed suicidal ideation, unknown long-term effects. When the regulator, the legislature, and the theologian all point the same direction, take the hint.


Ask your clients about it


Millions of people are already typing their worst days into chatbots, and some of your clients are among them. Ask directly, the same way you ask about substances or supplements. What a warm, endlessly available machine tells them between sessions is shaping your work whether you know about it or not. Give it a line in your intake paperwork and a place in your ongoing assessment.


Five actions for this month


  1. Pull your state's AI statute, or confirm your state has none yet, and calendar a six-month recheck.

  2. Update your informed consent to name any AI involved in your documentation or communication.

  3. Add a question about AI and chatbot use to your intake and your regular check-ins.

  4. Audit your tools. Anything touching protected health information needs HIPAA compliance and a signed business associate agreement.

  5. Draw the wall in your practice policies: AI on the administrative side, persons on the clinical side.


The law is drawing a line our doctrine drew centuries ago. The care of souls belongs to persons. Use the machines for the machinery of the practice, and keep the room for the image bearers.



Ready for a community that gets both the clinical and the spiritual side of this work? Join us at www.remnantcounselorcollective.com.

 
 
 

Comments


bottom of page