Three parts appear on the demo screen: (A) Demo Terms — accepted in order to use the demo; (C) Model-Training Notice — informs you, asks for no approval; (B) Model-Training Consent — an optional box; the demo works even if it is left unchecked.
A. Demo Terms
In short: What you say in the demo is turned into text and a report. Use the demo only with fictional cases or the sample cases we provide; do not enter real patient data. Your demo data is stored encrypted and deleted after at most 90 days; you can ask for it to be deleted earlier at any time.
- What these terms cover: All demos of AutoNote, Voice2Report, Dictation2Report, Voxy and all our future products offered via www.meditechlabs.ai or through access we provide. The demo is for presenting the product; it is not clinical decision support or a medical device and is not used in real patient care.
- Fictional cases only: The public demo was not designed for real patient data and may not be used for that purpose. Removing the name and ID number is not enough; a real patient's voice, case details or dates can make the person identifiable. If you want to try it with real patient data, use the institutional pilot route in item 3. If we notice that real patient data has been entered into the demo, we delete those records and remove them from model training.
- Institutional pilot and licensed use: Trials involving real patient data take place only under a pilot or licence agreement we have signed with your healthcare organisation, on the infrastructure your organisation chooses. In this use your organisation makes the decisions and is responsible for informing patients; we act on your organisation's instructions. This use is outside the public demo and is subject to the agreement, not to these terms.
- Which data we process: Your audio recording during the demo, the text and report generated from it, your browser and device information, the date and duration of the demo, and the identity and contact details you gave in the demo request form.
- What for: To provide the demo, fix errors and get back to you. Your demo data is used for model training only with your consent in section B.
- For how long: Raw audio recordings and outputs are kept for at most 90 days; situations that require earlier deletion (patient data entered by mistake, your deletion request) are reserved. If you gave model-training consent, only the de-identified training copy is kept for the period in section C.
- Where and how it is protected: Your demo audio and text data and your form data are processed and stored on our servers in Türkiye; they are not sent abroad. Data is encrypted in transit (TLS) and at rest; access is limited to our authorised employees and is logged. In live use, your organisation chooses one of the options: its own server, a local cloud, or the cloud specified in the contract.
- Intellectual property: The demo software and models belong to us; you may not use the demo for reverse engineering, copying or offering it to third parties. You may use the report output you generate for your own evaluation. You have no ownership, usage or compensation rights in the models trained with demo data (if you gave consent), their weights or derived outputs; these belong to us.
- Liability: The demo is provided "as is"; the accuracy of the outputs is not guaranteed and they may not be used for medical decisions.
- Your rights: Your rights and how to apply are explained in the Privacy Notice.
Legal basis: KVKK art. 5/2-c (pre-contractual negotiation) and art. 5/2-f (legitimate interest).
Box on the demo screen (required): "I have read the Demo Terms; I agree that I will enter only fictional or sample cases into the demo and no real patient data."
C. Model-Training Notice
In short: If you allow it, we improve our speech recognition and report models with your demo recordings. Recordings are de-identified before training, stay in Türkiye, are kept for at most 3 years and are not shared with anyone else. If you change your mind, they are not used in new training runs and are deleted.
| Heading | Description |
|---|---|
| Responsible party | MediTechLabs Yazılım A.Ş. |
| Which data | Your demo audio recording, the text and report generated from it, the demo date, your consent record (date, time, IP, text version, consent ID) |
| What for | To train, test and improve our speech recognition and report generation models |
| Fictional-case limit | The demo is for fictional cases only; we do not process your recordings as a real patient's health data. If we see that real patient data has been entered by mistake, we immediately block access to the record, remove it from training and delete it. Your consent covers only your own data; it does not create consent on behalf of patients. |
| De-identification | Before training, names, ID/file/protocol numbers, phone, e-mail, address, date of birth and full dates, and institution names are removed from the text; the corresponding parts of the audio are silenced; the audio is not used to recognise you by your voice. This process is not a guarantee of full anonymity; we continue to protect the recordings as personal data. |
| For how long | Raw recording at most 90 days; the de-identified training copy at most 3 years from the consent date, after which it is deleted or anonymised. Your consent record and the training records (containing no audio/text) are kept as evidence for 3 years after the training use ends; deletion logs are kept for at least 3 years from the date of the operation. |
| With whom | Our authorised employees and our hosting/backup providers established in Türkiye. Not shared with anyone else, not sent abroad. |
| If you change your mind | Just write to kvkk@meditechlabs.ai. Use in new training runs stops immediately; within 30 days the data is deleted from data sets, copies and backups. Training already completed is not repeated; however, we assess the risk of the model reproducing your data and, if needed, apply an output filter or retrain the model. |
| Your rights | Explained in the Privacy Notice. |
Legal basis: KVKK art. 5/1 explicit consent.
B. Model-Training Consent
I have read the Model-Training Notice above. I consent to the audio recordings created during the demo, and the text and reports generated from them, being processed by MediTechLabs Yazılım A.Ş. for the purpose of training, testing and improving its speech recognition and report generation models, under the conditions described there. I know that this consent is optional, that I can use the demo in the same way if I do not give it, and that I can withdraw it at any time. I declare that I have entered only fictional or sample cases into the demo and no data belonging to a real patient.
Box on the demo screen (optional, unchecked by default): "I have read the Model-Training Notice; I consent to my demo data being used for model training."
Consent record: When the box is checked, the date, time, IP, text version and a consent ID unique to you are stored together with your demo record in the consent store in Türkiye.