Oasis Ambient AI Series:
Part 2: Ambient AI Works - Here's How to Keep Compliant

Authors: Ashley Williams, Will Williams
Date: June 15, 2026

Ambient AI is changing how independent practices and rural health clinics operate. Providers who have deployed it successfully describe the same shift: they no longer focus on their laptops, and instead look patients in the eye. And, they leave work hours earlier. The technology delivers. Protecting that investment means understanding the regulatory framework around it.

Part 1 of this series explored how to build patient trust and manage AI refusals at the bedside. This part focuses on the compliance layer underneath: the federal software standards your vendor must meet and the state-level recording laws your practice must follow. Understand both layers to keep your AI implementation running and your practice protected.

Our team works with small, independent practices and rural health clinics nationwide, with a focus on the Mid-South: Arkansas, Louisiana, Mississippi, Alabama, Oklahoma, and East Texas. For practices in this region, navigating each state's recording requirements and transparency statutes is an essential operational responsibility. For practices outside these states, contact us, and we will help you identify your state's specific requirements.

An example of ambient AI: Azalea Health EHR's Clinical Assistant. 
Click Here for how Oasis Medical Solutions, an Azalea value-added reseller (VAR), helps practices implement secure AI.

A Brief Recap:
What is Ambient AI?

Ambient AI is a passive, intelligent listener. During a live patient visit, the software runs quietly in the background, captures the natural conversation between the clinician and patient, and converts it into a structured medical note. The provider reviews, edits, and signs every entry before it becomes part of the permanent record. The AI does not diagnose, prescribe, or make clinical decisions. It is a scribe.

Data from the American Hospital Association (AHA) shows that ambient tools save providers an average of 13 to 16 minutes per patient encounter—nearly 4 hours per day for a provider seeing 20 patients. That time goes back to patient care, not charting.

The Federal Baseline: Your Software Needs an "Off Switch"

The primary federal framework governing ambient AI in EHR systems is the Office of the National Coordinator for Health IT (ONC) HTI-1 rule, specifically the Decision Support Interventions criterion codified at 45 CFR § 170.315(b)(11).

In practice, this rule does not govern what happens in the exam room. It regulates your software. Under this standard, any certified EHR using predictive models or ambient documentation tools must do two things:

  • Show clinicians clear "source attributes" that explain how the AI model works and confirm that its outputs are free from algorithmic bias.
  • Allow authorized users to turn specific AI tools on or off at the point of care.

The second requirement is the one that matters most for daily operations. Federal law requires your EHR vendor to give you a functional "off switch" for AI modules. If your software makes it difficult to deactivate the ambient tool during a live encounter, it does not meet federal standards—and your practice inherits that compliance gap.

State-Level Enforcement: Know Your State's Rules

Federal HIPAA rules govern how patient data is stored, transmitted, and secured through Business Associate Agreements (BAAs). They do not govern the act of recording a conversation in the exam room. That is entirely a matter of state law.

Each state has its own wiretapping laws, medical board ethics standards, and established (or emerging) AI transparency requirements. Some states have enacted specific AI disclosure mandates with defined penalties. Others have no dedicated healthcare AI statute yet, but may still govern ambient tool use through existing informed consent and medical board rules.

The practical reality for practice managers is this: you cannot assume your state has no rules simply because there is no headline AI law on the books. Your state medical board likely has a position on informed consent and the use of emerging technologies in clinical settings. Your state's wiretapping statute defines who must consent before a conversation is recorded. Taken together, these existing frameworks create real obligations for any practice deploying ambient AI today.

The simplest approach eliminates the compliance question entirely: always ask the patient for permission before activating the ambient tool. This satisfies disclosure requirements in states where an active AI mandate exists, and protects practices in states where the law is still developing. Most importantly, it builds the kind of patient trust that makes ambient AI a long-term asset. It costs nothing and takes seconds.

Our team stays current on the specific requirements across our Mid-South focus region—Arkansas, Louisiana, Mississippi, Alabama, Oklahoma, and East Texas—and we actively monitor changes as state legislatures and medical boards continue to act. For a summary of your state's current requirements, contact us directly.

Is Your Software Ready?

Understanding the law only helps if your technology actually supports compliance. When a patient says "no," your EHR must allow the provider to deactivate the ambient tool immediately—without disrupting the encounter, restarting the system, or requiring the billing team to reconstruct the visit afterward.

Practices that use a standalone AI app disconnected from their primary EHR often run into this problem. The ambient note does not connect to the chart, consent logs do not carry forward, and the billing team loses the documentation trail they need to defend claims. Native integration between your AI tool and your EHR solves this directly.

Oasis Team Pro Tip: Let Your AI App do Double-Work

When a patient declines a recorded interaction, they are specifically withholding consent for the recording of that encounter, rather than restricting the provider's ability to use the software itself. Providers are still permitted to use the tool for dictation purposes after leaving the exam room. Using a tool like Azalea Health's Clinical Assistant, you can naturally speak your orders, assessments, and plans without all the "grammar and punctuation talk” traditional dictation software requires.

Key Takeaways for Practice Managers

  • * Verify the "Off Switch": Confirm that your EHR and your ambient AI vendor comply with the federal ONC HTI-1(b)(11) mandate. The software must allow providers to deactivate AI modules instantly at the point of care.
  • * One-Party Consent Is Not Full Compliance: Even in one-party states, AI transparency laws and medical board ethics require disclosure. Texas has an active mandate. Louisiana requires verbal disclosure. Mississippi requires written informed consent.
  • * Know Your State’s Rules: Even though states like Arkansas, Alabama, and Oklahoma have no dedicated AI healthcare statutes today, informed consent doctrine, HIPAA, and medical board ethics still govern practices. Patient communication and established consent procedures protect the practice from liability issues.

What's Next in the Series

Now that the trust gap (Part 1: Ambient AI and Patient Skepticism) and legal landscape are clear, how do you actually execute this in the exam room? In the rest of our series:

  • Part 3: Building a Transparent Patient Consent Workflow for Ambient AI (Release Date: June 2 2026)
    We walk through the exact scripts, front-desk workflows, and clinical pivots your team needs to handle patient opt-outs smoothly and professionally.
  • Part 4: Demystifying AI Security: Patient Myths and Vendor Vetting (Release Date: June 29, 2026)
    We explore how to address patient security fears with facts and what practice managers must look for when vetting AI vendors to ensure true compliance.

Resources & References


Who is 
Oasis Medical Solutions

Oasis Medical Solutions is a healthcare technology company that helps medical practices streamline their operations and maximize revenue. We specialize in providing Electronic Health Records (EHR) systems and related services, including practice management software, billing solutions, and consulting.

With a focus on personalized service and customized solutions, Oasis aims to empower healthcare providers to navigate the complexities of the healthcare industry and focus on delivering quality patient care.

Value-added reseller on a client call, viewing Oasis website and EHR software on a split screen, assisting with healthcare technology solutions.

State-Level AI & Recording Requirements: A Regional Reference

Texas
The Texas Responsible Artificial Intelligence Governance Act (TRAIGA), enacted as House Bill 149, has been fully active since January 1, 2026. Texas requires healthcare providers to deliver a "clear and conspicuous" disclosure in plain language to the patient on or before the date a service is first provided. The Texas Attorney General enforces this requirement, with financial penalties for non-compliance.

Reference: Becker's Hospital Review — 4 States Regulating AI in Healthcare

Louisiana
Louisiana has acted on two fronts. On May 28, 2026, the Governor received House Bill 475, which requires healthcare professionals to verbally disclose the use of any recording software to a patient before an AI-transcribed visit. Violations subject the provider to disciplinary action by their professional licensing board.

Separately, pending legislation (2025 House Bill 114) would establish a boundary between administrative and clinical AI use. If passed, it would prohibit AI from directly interacting with a patient or contributing to a diagnosis without licensed review and approval. Violations would carry civil penalties of up to $10,000 per incident.

References:
• Louisiana HB 475 — Patient consent prior to AI recording
• Louisiana HB 114 — AI use by healthcare providers (pending)

Mississippi
The Mississippi State Board of Medical Licensure classifies the "utilization of Artificial Intelligence" as an alternative medical practice under Part 2635, Chapter 13 of the Administrative Code (effective 2019). This is active regulation. Physicians must obtain written informed consent from the patient before using AI tools, and the licensed physician remains ultimately responsible for every clinical decision.

Reference: Mississippi State Board of Medical Licensure — Administrative Code, Part 2635, Chapter 13, Page 77

Arkansas
Arkansas is a one-party-consent state under Ark. Code § 5-60-120. No dedicated healthcare AI statute is currently in effect. Two AI-related healthcare bills introduced in the 2025 session (HB 1816 and HB 1297) were withdrawn by their authors before a committee vote. Existing medical ethics, informed consent doctrine, and HIPAA govern AI use in clinical settings.

Reference: PermitByState — AI Healthcare Regulations in Arkansas (2026)

Alabama
Alabama is a one-party-consent state under Ala. Code § 13A-11-30. No dedicated healthcare AI statute is currently in effect. No healthcare AI bills advanced in the 2025 or 2026 legislative sessions. Alabama's informed consent doctrine requires providers to disclose material information a reasonable patient would want to know. Whether AI involvement is considered material is an open question the state has not yet resolved by statute.

Reference: PermitByState — AI Healthcare Regulations in Alabama (2026)

Oklahoma
Oklahoma is a one-party-consent state. No dedicated standalone healthcare AI statute is currently in effect. SB 546, signed into law in March 2026 and effective 2027, establishes consumer data privacy rights but applies to large-scale commercial data processors — not individual HIPAA-covered healthcare practices. A prior healthcare AI bill (HB 3577) failed in 2024. Informed consent doctrine, HIPAA, and medical board ethics govern AI use in clinical settings.

References:
• Oklahoma House — SB 546: Oklahoma Consumer Data Privacy Act
• Recording Law — Oklahoma AI Laws and Regulation (2026)