Does Otter.ai use your recordings to train its AI?

Otter.ai's privacy policy says it trains its AI on de-identified recordings and transcripts. The policy, effective June 16, 2026, lists "training our proprietary AI technology on de-identified audio recordings and on transcriptions" as one of the purposes of its processing ¹. We found no clause in that policy that lets you opt out of this purpose. This answer is about the policy as we read it on October 17, 2026. Policies change, so check the current version yourself.

Otter's privacy policy names AI training as a purpose of your audio

The training purpose appears in the policy's section on how it uses your personal information, which lists it as one purpose among others ¹.

The parenthetical in that clause matters. The policy says the transcriptions "may contain Personal Information" ¹. So what the model learns from is not limited to the sound of your voice. The words you say can include names, phone numbers and addresses, and a transcript keeps those words.

"De-identified" is a written promise, and the policy does not describe the method

Otter says it will "maintain deidentified data in deidentified form and not attempt to reidentify the information, except solely as permitted for determining the effectiveness of our deidentification processes" ². That is a commitment, not a technical guarantee. The policy does not say how it removes identifying details from audio, and it does not say how well that works.

The policy also sets a limit for linked data. If Otter combines aggregated data with your personal information so that it can identify you, it "will treat the combined data as Personal Information" ³. That clause names aggregated data, not de-identified audio, so the policy does not say how it applies to a recording.

We found no opt-out from training in the policy

The rights section names two opt-outs. One covers the "sale" of personal information, and the other covers processing for targeted advertising ⁴. Neither names AI training. We found no clause that lets a consumer stop the training purpose on the policy page.

Two points remain open. Third-party guides describe a setting in the app, or a transcript-rating flow, that they say turns training off. We could not confirm either on Otter's own pages. The policy also does not say which plans the training purpose covers, and we did not verify Otter's business or enterprise terms.

If the answer matters for your account, ask Otter in writing which setting, if any, stops training on your recordings. Keep the reply with its date. A written answer from the company is more reliable than a summary from a guide.

A class action says recording consent was missing, and its claims are untested

A federal class action filed in August 2025 alleges that Otter Notebook, by default, does not ask meeting attendees for permission to record ⁵. The complaint also says participants are not told that recordings are shared with Otter to improve its AI systems. These are allegations in a complaint. We did not find a court ruling on them, and we did not check the current status of the case.

Consent and training are separate questions. Consent asks whether the people in the room knew they were being recorded. Training asks what the company does with the recording afterwards. If you record other people, you need answers to both.

What to do if you use Otter or a similar meeting tool

1. Open the privacy policy and search it for "train". Note the version date at the top of the page.
2. Tell participants that you are recording, and that a transcription service will process the audio. Say whether the service may use it to train its AI.
3. Ask the vendor in writing which setting, if any, stops training on your plan, and keep the reply.
4. If you do not want your audio used for training, use a tool that recognizes speech on your device. Our app, Yap, which we make, runs recognition on your computer. Its settings and policy still decide what you share.

For the settings in other apps, see how to stop apps from training on your voice. For the risk that a trained model repeats what it heard, see whether AI models can leak what you dictated. The AI dictation and voice profile entries define two terms used here. For the wider comparison, read our guide on whether dictation apps train AI on your voice, or the voice privacy overview.