Legal
AI Disclosure + Acceptable Use
Last updated May 23, 2026 · Effective immediately
Health-E-Now uses AI to make virtual care faster and less administrative — never to replace your physician. This page explains exactly where AI is in the loop, what it sees, and how to use the platform without it.
Where AI is in the workflow (in addition to the table below)
- OCR on government-issued ID at signup. When you upload a government-issued ID as part of Stripe Identity verification, optical character recognition pre-fills your account with the name, date of birth, and document expiry from the ID — so you don't re-type them. The biometric face-match is performed by Stripe; we only receive a pass/fail result and a verification session ID, not the ID image itself.
What AI does not do
- AI does not diagnose you. Diagnosis is the physician's responsibility.
- AI does not prescribe medication. A draft prescription, if generated, is only valid once a CPSO-licensed physician reviews and electronically signs it.
- AI does not recommend treatment. Plans and instructions come from the physician.
- AI does not decide whether to refer you for in-person care, urgent care, or emergency care. The physician makes that call. The chatbot also routes obvious emergencies (chest pain, stroke symptoms, suicidal ideation) directly to 911 or 9-8-8 before a visit is booked.
Where AI is in the workflow
| Surface | AI vendor | What it sees | What it produces |
|---|---|---|---|
| Pre-visit intake chatbot | Anthropic Claude (Sonnet) | Your answers to a guided conversation about your symptoms, history, and what you're hoping to get from the visit. | A draft summary attached to your chart for the physician to review before the consult. |
| Live transcription during the visit | Deepgram | The audio of your visit, streamed in real-time. Audio is not stored beyond the visit unless you separately consent to recording. | Live captions and a transcript used by the AI scribe. |
| Chart note drafting (AI scribe) | Anthropic Claude (Sonnet) | The visit transcript and the physician's edits. | A SOAP-style draft (Subjective, Objective, Assessment, Plan) that the physician edits and signs. |
| Document drafting (Rx, sick notes, lab requisitions, referrals) | Anthropic Claude (Sonnet) | The physician's plan and patient details required for the document. | A draft document for the physician to review. |
Physician responsibility
The treating physician is the Health Information Custodian under PHIPA for the medical record they create. They are solely responsible for the content of any prescription, requisition, sick note, referral, or chart note they sign — whether or not AI produced the draft.
Where your data goes
Anthropic (Claude) and Deepgram (transcription) are US-based service providers. We use both under written agreements that prohibit training on our data, require zero retention beyond the request lifecycle, and route to North American infrastructure. Because they are US providers, requests may be subject to US lawful-access processes; we disclose this on our Privacy Policy. PHI at rest stays in Canada (Supabase Toronto / AWS ca-central-1).
Opting out
- Decline AI intake. Before any visit, you can choose "skip the AI intake" from the booking screen. A staff member will walk you through the same questions over text or audio.
- Decline AI scribe / transcription. The consent modal at visit start lets you decline. Your physician will type their own notes, and no audio is streamed to Deepgram or Claude.
- Withdraw consent later. From Account → Privacy, you can flip "AI scribe" off for all future visits.
Declining AI does not affect your care. You see the same physicians, on the same platform, on the same schedule.
How we keep AI honest
- Every AI-drafted note is read and edited by the treating physician before it's signed. The signed document is the physician's, not the AI's.
- Physician edits over time tell us where the AI is weak. We review override rates monthly and tune prompts in response — without ever overriding clinical judgement.
- Patients can request a copy of any AI-generated draft attached to their visit through Account → Privacy → Download my data.
Acceptable use
Don't try to use the chatbot for non-medical purposes (e.g., as a general assistant, to generate content unrelated to your care, or to bypass the scope-of-care limits in our Terms of Use). The chatbot is designed for clinical intake; off-topic prompts will be ignored.
Questions
Email our Privacy Officer at privacy@health-e-now.com or visit our Complaints + inquiries page.