Does Oura train on your data?
Oura's policy puts its own use of your data under a heading that says "Large Language Model Training", and rules out only selling it for somebody else's LLM. Whether it trains a model, it never says.
This clause is real, but it doesn't resolve the training question - see the finding below.
"Large Language Model Training. We may collect and use personal data to develop, test, improve, and support Oura's own AI, machine learning, and large language model-powered features. We do not sell personal data for the purpose of training third party Large Language Models ("LLMs")."
- 7 Sept 2026
- 5 Oct 2026
- All accounts (Oura Ring and subscription)
- global
- Fetched from the company
- 2953163F…35BF
Why this is still UNCLEAR: Re-read on 12 September 2026 against the archived copy; the first reading described only the second of the paragraph's two sentences. The heading says Large Language Model Training. The sentence under it says personal data may be used to "develop, test, improve, and support Oura's own AI, machine learning, and large language model-powered features"; the sentence after it rules out selling personal data to train third-party LLMs. So the company files its own use under a training heading, and elsewhere promises pseudonymisation, aggregation or de-identification when it develops or refines AI features — the measures applied to material a model is built from. What the document never says is that your heart rate, sleep stages or temperature train a model: the object of every sentence is a feature, not a model, and "train" is used of Oura's own systems nowhere. Training named, then qualified — which is where this entry sits on the patterns page.
What each plan can refuse
The same questions, asked of every plan we could identify. Every cell comes from a quoted clause or a documented search - see the evidence above.
- Not established
- Not stated
- —
- —
- Not established
How to opt out of Oura AI training
Plan by plan, from what Oura's own policy says about refusing. Where it publishes the steps, they're listed here with the limits they carry. Where there's nothing to switch off, or nothing graded, this says so instead of inventing a menu path.
All accounts (Oura Ring and subscription)
Not graded, so no route is pointed at. Re-read on 12 September 2026 against the archived copy; the first reading described only the second of the paragraph's two sentences. The heading says Large Language Model Training. The sentence under it says personal data may be used to "develop, test, improve, and support Oura's own AI, machine learning, and large language model-powered features"; the sentence after it rules out selling personal data to train third-party LLMs. So the company files its own use under a training heading, and elsewhere promises pseudonymisation, aggregation or de-identification when it develops or refines AI features — the measures applied to material a model is built from. What the document never says is that your heart rate, sleep stages or temperature train a model: the object of every sentence is a feature, not a model, and "train" is used of Oura's own systems nowhere. Training named, then qualified — which is where this entry sits on the patterns page.
What this policy has done since we started watching
No change detected. We've held a copy of Oura's policy since 30 Aug 2026 and re-read it since, across 2 captures. The text hasn't moved in a way that touches this verdict. That's a fact about the document, and it's only knowable by having kept the earlier copies.
Everything we read
Vendor response
We haven't yet written to Oura about this verdict, and they haven't disputed, clarified or announced a change to any clause quoted above. When either happens it's published here, word for word.
Report a wrong clause