Welcome to this week’s Dental AI Weekly, honest analysis of where dental AI is going, from someone building in it.
AI disclosure got more specific this week.
Louisiana now requires verbal disclosure before recording an appointment for AI transcription. The FDA opened a public discussion on how generative AI-enabled medical devices should be evaluated and monitored. And a new patient-trust commentary pulled the same issue into plain language: patients are not just reacting to the word AI. They are reacting to uncertainty about privacy, responsibility, and whether a human is still in the loop.
That is the thread running through this issue. The next phase of dental AI will not be won by the flashiest feature demo. It will be won by tools and teams that can show what the software does, where the output goes, and where human review still controls the final decision.
WHAT HAPPENED THIS WEEK
01 – BEFORE RECORDING
Louisiana now requires a verbal disclosure before an appointment is recorded for AI transcription.
Phelps Dunbar reported on August 21 that Louisiana Act 649 took effect August 1. The law requires a health care professional to verbally disclose the use of a recording device, software, or service before recording any part of an appointment or treatment that will be transcribed by AI.
The same legal analysis explains that Texas has separate notice requirements when AI is used for diagnostic purposes or treatment. The rules differ by use and state. A script that fits Louisiana’s transcription rule may not cover Texas’s diagnostic or treatment notice.
What this means for your practice: Disclosure is becoming an operating workflow, not a footnote in intake paperwork. The practice that treats AI consent like a normal chairside habit will be in better shape than the practice that waits until a patient notices the microphone.
02 – PATIENT TRUST
A new commentary tied patient trust in AI to privacy, accountability, and access to a person.
A Journal of Medical Internet Research commentary published August 21 responded to a systematic review of 25 qualitative studies on patient concerns about health care AI. Across those studies, recurring concerns included privacy, data security, opaque decision-making, damage to the clinician-patient relationship, unclear accountability, equitable access, and replacement of human workers.
The commentary recommends informed-consent guidance, clinician review checkpoints, and clear responsibility among providers, health systems, and AI vendors. The recommendations are the author’s interpretation of the review. The commentary itself was not peer reviewed.
What this means for your practice: Trust is becoming part of AI performance. If the patient experience feels automated, vague, or hard to challenge, the tool may save time while costing confidence.
03 – DIFFERENT JOB, DIFFERENT PROOF
The FDA opened a discussion on how generative AI medical devices should be evaluated.
On August 18, the FDA asked for public feedback on risk assessment, premarket evaluation, postmarket monitoring, and other regulatory questions for generative AI-enabled medical devices. Feedback is due October 19 under docket FDA-2026-N-7874.
This is a discussion paper, not draft guidance, final guidance, or a policy change. It applies to generative AI-enabled medical devices, not every scribe, scheduling tool, or administrative workflow.
What this means for your practice: The evidence bar should match the job. A scribe draft, a diagnosis aid, and a treatment recommendation should not be bought, trained, or supervised as if they carry the same risk.
“The evidence bar should match the job.”
BY THE NUMBERS
772
large health care data breaches affecting 500 or more people were listed for 2025 on the HHS Office for Civil Rights breach portal as of June 2026.
What it signals: This is the wider backdrop for this week’s AI news. Patients are bringing breach anxiety into any tool that records, stores, or summarizes health information. A practice that can describe its recording workflow in plain English will feel different than one that sounds surprised by the question.
Source: HIPAA Journal, June 5, 2026, based on the HHS OCR breach portal. The total is health care-wide, not dental-only or AI-specific, and may change as investigations conclude.
READER Q&A
“A patient asked whether AI was listening during the appointment. What should my team explain before we start recording?”
Practice question selected for Issue 027
BH: If the patient has to ask whether AI is listening, the practice probably did not set the expectation clearly enough.
Before recording begins, tell the patient what the technology does and why you use it. Keep the explanation simple:
“We use an AI scribe to listen during the appointment and help us prepare your clinical notes. Your provider reviews the note before it becomes part of your chart. Is it okay if we use it during your visit?”
Give the patient a real opportunity to say no, and make it clear that declining will not affect their care. Your team should also be ready to explain how recordings and patient information are handled, who can access them, and whether the audio is retained.
Include AI-assisted recording in your privacy and consent policies, but do not rely on a form the patient may have signed years ago. Notices at the front desk, in the lobby, or in each operatory can reinforce the policy. They do not replace the chairside conversation.
In our experience, most patients give a simple yes. Some are curious and want to know more. Either response is easier to handle when the practice explains the technology before pressing record instead of waiting for the patient to notice.
Recording and consent requirements vary by state, so practices should have their policies and scripts reviewed by qualified counsel. Regardless of the legal minimum, clear disclosure and affirmative consent are the better standard for patient trust.
FROM THE ORACORE BLOG THIS WEEK
Privacy | 3 min
Ambient AI Dental Privacy: Consent, BAA, and Review
A practical review of what is captured, who can access it, how patients are informed, and who approves the output.
Compliance | 3 min
Dental AI Privacy and Consent Guide
A short buyer’s guide to business associate terms, recording policy, minimum necessary access, data retention, and state-law review.
OraCore Scribe gives the provider a reviewed-note workflow, not a black box. Capture the visit, review the draft, and finish documentation with the clinician still in control.
That’s the week. Reply with what you’re seeing in your own practice. I read every one.
Brad Hutchison
CEO, OraCore AI