Does Hotels.com train on your data?
Hotels.com names fourteen things it uses AI for, then adds "Other". Whether your data trains the models behind them is the one use it does not name.
This is not silence, which is what makes it worth reading. The statement carries a section headed "Our use of artificial intelligence" opening "We use artificial intelligence (AI) and machine learning (ML) for various purposes to deliver our platform and associated services", and then names fourteen of them — pricing, fraud, feature generation, auto moderation, chatbot, insurance, search engine optimisation, content generation, search and ranking, recommendations, security governance, productivity, anomaly detection and images — before adding a fifteenth headed "Other", for uses "not captured by the above categories". Having set out that much, it never says whether any of your data is used to build the models behind any of it. The nearest approach is "To enrich other applications such as embeddings", which describes something derived from your data without saying it is training. One phrase does recur throughout the purposes table — "Market research, analytics and training purposes to improve our services", 55 times — and it is the reason this entry cannot simply record an absence: it neither says model training nor rules it out, and the same document's only unambiguous use of the word is a call-recording clause about "quality control, training, dispute resolution". Fourteen uses named, and the fifteenth question unanswered. This entry originally said "machine learning does not appear" and counted ten applications. Both were wrong: the section is rendered in spans inside leaf divs and this site's text extractor was dropping them. The extractor was fixed on 1 September 2026 and the cited capture re-read.
- Hotels.com Privacy statement, last updated 10 December 2025, saved from the live site on 2026-08-30 and re-read on 2026-09-01 with the corrected text extractor (https://uk.hotels.com/legal/privacy)
- train, training, fine-tune, machine learning, artificial intelligence, AI model, large language model, embedding
- 0 of 63 raw hits
- 1 Sept 2026
"Training" accounts for 57 of these, and 55 of those are a single phrase the purposes table restates for each category of data it describes: "market research, analytics and training purposes to improve our services". The remaining two senses are unambiguous and not about models — a call-recording clause covering "quality control, training, dispute resolution", and a commitment to "provide regular information security training for employees". "Machine learning" appears twice and "artificial intelligence" four times, in the AI section's heading and opening sentence. None of them is a statement about what trains a model, which is why the count above is zero despite an entire section on AI.
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 Hotels.com AI training
Plan by plan, from what Hotels.com'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
Not graded, so no route is pointed at. This is not silence, which is what makes it worth reading. The statement carries a section headed "Our use of artificial intelligence" opening "We use artificial intelligence (AI) and machine learning (ML) for various purposes to deliver our platform and associated services", and then names fourteen of them — pricing, fraud, feature generation, auto moderation, chatbot, insurance, search engine optimisation, content generation, search and ranking, recommendations, security governance, productivity, anomaly detection and images — before adding a fifteenth headed "Other", for uses "not captured by the above categories". Having set out that much, it never says whether any of your data is used to build the models behind any of it. The nearest approach is "To enrich other applications such as embeddings", which describes something derived from your data without saying it is training. One phrase does recur throughout the purposes table — "Market research, analytics and training purposes to improve our services", 55 times — and it is the reason this entry cannot simply record an absence: it neither says model training nor rules it out, and the same document's only unambiguous use of the word is a call-recording clause about "quality control, training, dispute resolution". Fourteen uses named, and the fifteenth question unanswered. This entry originally said "machine learning does not appear" and counted ten applications. Both were wrong: the section is rendered in spans inside leaf divs and this site's text extractor was dropping them. The extractor was fixed on 1 September 2026 and the cited capture re-read.
What this policy has done since we started watching
No change detected. We've held a copy of Hotels.com's policy since 1 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 Hotels.com 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