Here’s the sentence your AI policy needs this month: no customer, client or colleague health information goes into any AI tool unless it’s on a plan the company has bought and reviewed. ChatGPT Health is why. Its privacy promises protect the person who signs up, for that person’s own data. They do nothing for the client whose sick-leave note an employee pastes in.
On January 7, 2026, OpenAI opened a waitlist for ChatGPT Health, a separate space inside ChatGPT for health questions, medical records and fitness data. The promises are made to one user about that user’s information, and your company isn’t a party to them.
Key points
- OpenAI says Health has separate storage and memories, added encryption, optional multi-factor sign-in and no foundation-model training on Health conversations.
- The launch text names Free, Go, Plus and Pro users, leaves Business and Enterprise unnamed, and doesn’t mention HIPAA.
- Canada has no standalone AI law; privacy law carries the main penalties, and its reach here varies by province.
- One written sentence, one named person to ask, and a check of the AI plans you pay for close the gap.
What OpenAI’s text promises, and to whom
OpenAI says Health runs as its own space with separate storage and memories, so health details stay out of ordinary chats. Users can view or delete Health memories and disconnect apps at any time. Connectors include Apple Health, Function, MyFitnessPal, Weight Watchers, AllTrails, Instacart and Peloton. Medical records come in through b.well, a U.S. health data connectivity company, and OpenAI limits that part to U.S. users. On July 23, 2026, a note on the page described a wider launch to U.S. adults on web and iOS.
Read the text for what it omits. OpenAI’s launch text names consumer tiers only and doesn’t mention the U.S. federal health privacy law. We don’t call either a flaw. A consumer health tool and a regulated health service are different products, and the line between them is drawn in contracts and statutes. Product copy doesn’t draw it.
So the promises govern what OpenAI does with what a user puts in their own space. They say nothing about a team lead pasting an employee’s accommodation note into a chat, or a bookkeeper uploading a client’s medical receipt for a summary. A feature description is no data processing agreement, and no firm can hold up a settings screen to a regulator as due diligence. Vendor assurances count when they sit in a written agreement on a plan you control.
A case we made up: a 40-person benefits consultancy, an adviser with a personal Plus account, a client’s sick-leave note pasted into Health to draft a reply. The client agreed to nothing with OpenAI, and neither did the firm. If the client asks where the note went, “the employee’s settings” is a thin answer.
We haven’t tested Health. Everything above is OpenAI’s own description.
Where Canadian privacy law sits
Canada has no standalone AI law, and as we’ve explained, privacy law carries the main legal penalties for AI use. How regulators would treat staff health details typed into a personal ChatGPT account probably varies by province and by who holds the data. Quebec sets its transfer test at the provincial border, which our look at PC Chat’s French launch in Quebec walks through. We aren’t lawyers, and the question has no tidy answer.
That uncertainty argues for a written rule. A firm that can say “we told staff exactly what not to paste, in writing, on this date” stands far better than one relying on a vendor’s settings page.
Writing the rule, and the sceptic’s objection
Add the sentence to whatever AI or confidentiality policy you have. Staff may use AI health tools for their own health on their own accounts. Anyone unsure asks before pasting, and one named person answers. Name health explicitly: it turns up in HR files, benefits claims, return-to-work notes and customer service threads, and “sensitive information” is easier to miss. Our guide to privacy safeguards for AI email workflows shows how short such a rule can be. Put it in the onboarding pack, because new hires bring their own habits and accounts.
Then list the AI plans the company pays for, such as those in our Microsoft Copilot plan and pricing guide, and read whether each vendor’s terms address health data. If ten minutes doesn’t find the answer, that’s your answer.
The sceptic says this is a fuss over a consumer feature. People already ask chatbots about symptoms, and a dedicated space with separate memory beats a general chat. Agreed, for the individual. The firm holding someone else’s information gets nothing from it, and an employee with a personal account and a colleague’s note is the case that ends up in an HR complaint. We’d soften our view if OpenAI put the same commitments in writing for business plans, or if independent reviewers confirmed the isolation and training claims. Until then, send the sentence to staff this week.
Frequently asked questions
Does ChatGPT Health use my health chats to train OpenAI’s models?
OpenAI says it does not use Health conversations to train its foundation models. That is the company’s own statement, so treat it as a claim until independent reviewers check it.
Can Canadians use ChatGPT Health?
At launch OpenAI named users outside the EEA, Switzerland and the UK, which would include Canada. The medical records import was limited to the U.S., so the feature set is narrower here.
Should employees put customer health information into ChatGPT?
Only on a plan your company has bought and reviewed, and never on a personal account. Add a single sentence to your AI policy so the rule is written down.
Archive entry dated January 8, 2026, written and fact-checked on October 8, 2026. Sources are linked on the claims they support.