Is ChatGPT HIPAA compliant?
Last updated: November 17, 2026
The short answer
Only some ChatGPT products, and only with a signed agreement. OpenAI signs a Business Associate Agreement (BAA) for a short list of products. Without a BAA for the exact product you use, you should not put protected health information (PHI) into it. The everyday ChatGPT plans most people use are not on that list.
Which OpenAI products can be used with a BAA
As listed by OpenAI's Help Center in November 2026:
- ChatGPT for Healthcare
- ChatGPT for Enterprise with Regulated Workspace
- ChatGPT for Clinicians
- ChatGPT FedRAMP
- The OpenAI API with Modified Retention (and API FedRAMP)
ChatGPT Free, Plus, Pro and Business do not appear on OpenAI's list of HIPAA-eligible products. Even on an eligible product, OpenAI notes that a feature being available does not by itself mean that use is covered by your BAA.
A BAA is the start, not the finish
OpenAI's HIPAA guide puts several duties on you, the customer. Among them: manage access with single sign-on, don't put PHI in support requests or screenshots sent to OpenAI, and don't use the service to keep PHI as part of a designated record set. You still need your own risk analysis, staff training and access controls.
What if your team uses regular ChatGPT anyway?
Many clinics, billing teams and health-tech startups have staff on personal or Business plans. The safest rule is simple: patient details never go into those tools. In practice that means removing names, record numbers and other identifiers before anything is pasted or uploaded.
Omna does this on each person's computer, before anything leaves. It masks names, contact details, dates of birth, government IDs, clinic and hospital names, and labelled health details such as medical record numbers, patient IDs, lab accession numbers, NPIs, insurance IDs and diagnosis codes — replacing each with a placeholder like [HEALTH_1]. The AI only ever sees the placeholder.
Be clear about the limits: Omna hides who the patient is, not the medical facts. An illness, a medicine, a lab result or the date of a visit written in plain words is not hidden, so the AI can still help — and some files are sent as they are. Omna says so every time. See what Omna hides and where masking stops.
Next steps
- Write down which AI tools your staff may use — start from our free AI acceptable use policy template.
- See what proof of safe AI use looks like in the sample team report.
- Omna never receives your data — masking happens on each person's own computer, so patient details never reach Omna. If your compliance team needs a BAA or other paperwork, talk to us.
Sources
- OpenAI Help Center — HIPAA eligible products and functionality
- OpenAI Help Center — Getting a Business Associate Agreement for the OpenAI API
- OpenAI — HIPAA Implementation and Configuration Guide
- OpenAI — Business data privacy, security, and compliance
Note: this page is general information, not legal advice. OpenAI changes its plans often — check its current list before relying on it. Omna is a privacy tool, not a compliance certification.