AI use agreement

1.The key points

The six points below summarise this agreement. The sections that follow explain them in full.

1. 1The AI features (AI scribe, AI consult, AI chat, document summaries, pre-visit briefs, the voice assistant and suggested actions) produce suggestions generated automatically by third-party AI models. They can be incomplete, outdated or wrong.
2. 2AI output is not a diagnosis, not medical advice and not a prescription. Clinova is not a medical device.
3. 3You must review and verify every AI output before using it. You remain solely responsible for every clinical decision, note, prescription and message made with it or after it.
4. 4Never rely on AI output in an emergency.
5. 5The clinic and its users, not Clinova, are responsible for how AI output is used with patients.
6. 6The patient context sent to the AI providers is pseudonymised, as described in the Privacy Policy.

2.Scope and acceptance

Doctors, nurses and clinic managers accept this agreement before their first use of an AI feature, and again whenever it changes, by ticking the acceptance box and typing their full name as an electronic signature. Clinova records the version accepted, the name typed, the date and time and the IP address.

The clinic makes sure that the members of its team who use the AI features follow this agreement.

3.What the AI features are

Clinova offers optional features that use artificial intelligence:

• the AI scribe, which turns a recorded or dictated visit into a draft note;
• AI consult, which offers clinical decision support on the patient file;
• AI chat, which answers questions about a patient file;
• summaries of uploaded documents and pre-visit briefs;
• the voice assistant, which answers spoken questions and prepares actions;
• suggested actions, such as a follow-up, a medication or a note to add to the patient file.

Their output is generated automatically by AI models run by third-party providers, from the information in the patient file and from what the user says, types or uploads.

4.The limits of AI

AI models can write text that sounds convincing but is incomplete, outdated or wrong. They can mishear a recording or misread a document, miss information that is in the file, confuse details when several patients are discussed, invent references, values or doses, and give different answers to the same question. Their output may reflect biases present in the data they were trained on.

The AI does not examine the patient. It knows only what is in the record and in the request, and it does not know what happened after the information was recorded.

5.Not medical advice, not a medical device

AI output is not a diagnosis, not medical advice and not a prescription. Clinova is not a medical device and has not been certified as one by any authority.

The AI features are writing and information aids for health professionals. They are not intended to diagnose, treat, cure or prevent any disease, and must never be the sole basis of a clinical decision.

6.Human oversight and verification

Human oversight is mandatory. When you use an AI feature, you agree to:

• read every AI output and check it against the patient file, your own examination and current clinical guidance before you use it;
• correct or delete anything wrong or unsupported before a draft becomes part of the record, a prescription or a message;
• confirm doses, interactions, contraindications and allergies from a reliable source;
• never let the AI make a decision on its own, and never send AI output to a patient without reading it first.

Changes that the AI proposes to the patient file, such as a medication, an allergy or a follow-up, are applied only after a user confirms them. By confirming a proposal, you confirm that you have checked it.

7.No use for emergencies or triage

Never rely on the AI features in an emergency, to decide how urgent a patient’s condition is, or whenever time is critical. They can be slow, unavailable or wrong. In an emergency, follow your clinical judgement and your emergency protocols, and call the emergency services where appropriate.

8.Clinical responsibility

You remain solely responsible for every clinical decision, note, diagnosis, prescription, order and message that you make with the help of an AI feature or after using one, exactly as you would be without it.

The clinic decides who in its team may use the AI features, and is responsible for their training and supervision and for how AI output is used with its patients. Clinova is not responsible for how the clinic or its users use AI output.

9.Informing patients and recordings

Where the law or professional rules require it, you inform patients that AI assisted their care or its documentation, and you obtain the patient’s consent before recording a visit for the AI scribe or using the voice assistant during a visit. The clinic sets the rules its team follows.

10.Only the data that is needed

Do not type, dictate or upload information the AI does not need for the task, such as a patient’s contact details, identity documents or details about other people.

11.AI providers and patient data

The AI features are run by third-party providers: Anthropic (Claude) and Microsoft Azure OpenAI write the text, and OpenAI transcribes recordings. They process the data under their own terms for business customers.

The patient context sent to them is pseudonymised. The patient is described by a reference, age and sex, without name, date of birth or contact details, and Clinova removes the patient’s name, e-mail addresses and phone numbers from free-text notes before sending them. The voice assistant is the one exception: to find the patient named aloud, it shares patient names, but never dates of birth or contact details. Clinova does not use patient data to train AI models.

Read the Privacy Policy→

12.Accuracy, availability and AI allowance

Clinova does not guarantee that AI output is accurate, complete, up to date or suitable for a particular patient, nor that the AI features are available at any given time. They depend on the providers, whose models change as they are updated.

AI use is limited by the clinic’s monthly AI allowance. When it runs low, the AI features may switch to a lighter model or stop until the next period or a top-up.

13.Liability

To the fullest extent permitted by the applicable law, Clinova is not liable for any loss, harm or claim that results from using or not using AI output, including any clinical decision, record, prescription or message based in whole or in part on it, nor for errors or omissions in that output. The clinic remains responsible towards its patients for the care they receive, including care in which AI output was used.

The limitations of liability in the Terms of Service also apply. Nothing in this agreement limits a liability that the law does not allow to be limited.

14.Breach of this agreement

If a user breaches this agreement, for example by using AI output without checking it or by entering data that is not needed, the clinic or Clinova may suspend that user’s access to the AI features. In case of serious or repeated breach, Clinova may suspend the AI features of the clinic.

15.Records and changes

Acceptance records can be seen by the clinic’s administrators and are kept as described in the Privacy Policy.

When this agreement changes in a way that matters, you are asked to accept the new version before you continue. Until you accept it, the AI features are not available to you.