What Happens to Your Audio? Dental AI Privacy Explained

It is the first question any sensible clinician asks about ambient documentation, and it is the right one: if the appointment is being listened to, what happens to the recording? The answer is the difference between a tool you can trust with patient conversations and one you cannot.

Processed, not kept by default

The sound of the appointment is used to produce the transcript and then the note. By default the audio is not stored afterwards: the deliverable is the clinical note, not an archive of patient conversations sitting on a server. Audio is only ever retained if you actively opt in, for example where you have a specific reason to keep a recording. The default, and the safe position, is that nothing is kept once the note is made.

This distinction matters because the two designs carry very different risk. A tool that quietly retains every recording accumulates a large and sensitive store of patient data. A tool that discards the audio once the note is produced holds almost nothing.

Why this design is the safe one

Stored recordings are a liability. They are sensitive, they accumulate over time, and they widen the surface for anything to go wrong, whether through breach, misuse, or simple loss. Not retaining them by default removes that risk at the source. The less patient data a tool holds, the less there is to protect, and the smaller the consequences if something does go wrong. This is data minimisation in practice, and it is the principle a well-designed clinical tool should follow.

What this means for patients

You should still tell patients that a tool is being used to help write the notes, in the same spirit of openness you bring to anything else in their care. Transparency is part of trust, and most patients are reassured to hear that the recording is not kept by default. If you do opt in to retaining audio for a specific purpose, factor that into what you tell patients and into your own data handling, since you are then holding sensitive data that you were not before.

Questions worth asking any vendor

• Is audio stored by default, and if so for how long?
• If retention is optional, what triggers it and who controls it?
• Where is data processed, and under what safeguards?
• What registrations and security standards does the vendor hold?

A trustworthy answer to "what happens to the audio?" should be short, clear, and reassuring. Evasiveness or vagueness on this question is itself informative.

A worked example

Two tools both produce good notes. One keeps every recording indefinitely "in case it is needed"; the other discards the audio once the note is written, unless you opt in. A year on, the first holds thousands of recorded patient conversations, every one of which is sensitive data that must be secured, governed, and accounted for. The second holds none of that. If a breach ever occurs, the difference between the two is enormous. The note quality might be identical; the risk profile is not.

Frequently asked questions

Is my patients' audio stored?

By default, no. It is processed to produce the note and then not retained, unless you opt in to keep it.

Why is not storing audio safer?

Because retained recordings are sensitive data that accumulate and must be protected. Discarding them removes that risk at the source.

Do I need to tell patients?

You should be open that a tool helps write the notes, as with anything in their care. Most patients are reassured the recording is not kept by default.

What if I opt in to keep audio?

Then you are holding sensitive data you otherwise would not, so reflect that in your patient communication and data handling.