Does Amazon train on your data?
Amazon's privacy notice runs to 28,000 characters and does not use the words "train", "AI" or "machine learning" once.
Amazon sells an AI assistant, runs one of the largest model-training businesses in the world, and holds the purchase history of a substantial share of the country. Its consumer privacy notice contains no form of "train", no "machine learning", no "artificial intelligence" and no bare "AI". Whether any of what it lists is used to build a model is not a question this document takes up. Two things a careful reader will find. The nearest language in the notice is a legitimate-interests clause covering "when we use your voice, video, or camera input to improve services" — improvement of services from Alexa and Ring-class input, which is the boilerplate this site declines to grade on, and which never names a model. And the notice links onward, in a list of further notices, to a "Generative AI Development Disclosure". Amazon's help pages answer 403 or 503 to every automated reader this site has, and the Internet Archive holds no copy, so the disclosure was saved by hand on 10 September 2026 and is not re-read automatically. It is about the datasets behind Amazon's generative AI services: "licensed and proprietary datasets, synthetic datasets, open-source datasets, and publicly available content (including web crawled data)", which "may contain public domain content, rights-protected material, and in some cases, personal information or aggregate consumer information". Whether your orders, searches or recordings are among the proprietary data it trains on is the one thing neither document says, so the finding stands.
- Amazon privacy policy, captured 2026-08-22 (https://www.amazon.co.uk/gp/help/customer/display.html?nodeId=GX7NJQ4ZB8MHFRNJ) · Amazon Generative AI Development Disclosure, saved by hand 2026-09-10 (https://www.amazon.co.uk/gp/help/customer/display.html?nodeId=TmGoGN3UbFaQ1CAph7)
- train, training, trains, fine-tune, machine learning, artificial intelligence, AI model, large language model
- 10 of 18 raw hits
- 10 Sept 2026
Every match is in the disclosure; the privacy notice has none. They describe what the generative AI services are trained on — licensed, proprietary, synthetic, open-source and publicly available datasets — and the safeguards around that training. None says whether the personal information Amazon holds about its own customers is in those datasets, which is the question this entry asks, so the matches do not resolve it.
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 Amazon AI training
Plan by plan, from what Amazon'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. Amazon sells an AI assistant, runs one of the largest model-training businesses in the world, and holds the purchase history of a substantial share of the country. Its consumer privacy notice contains no form of "train", no "machine learning", no "artificial intelligence" and no bare "AI". Whether any of what it lists is used to build a model is not a question this document takes up. Two things a careful reader will find. The nearest language in the notice is a legitimate-interests clause covering "when we use your voice, video, or camera input to improve services" — improvement of services from Alexa and Ring-class input, which is the boilerplate this site declines to grade on, and which never names a model. And the notice links onward, in a list of further notices, to a "Generative AI Development Disclosure". Amazon's help pages answer 403 or 503 to every automated reader this site has, and the Internet Archive holds no copy, so the disclosure was saved by hand on 10 September 2026 and is not re-read automatically. It is about the datasets behind Amazon's generative AI services: "licensed and proprietary datasets, synthetic datasets, open-source datasets, and publicly available content (including web crawled data)", which "may contain public domain content, rights-protected material, and in some cases, personal information or aggregate consumer information". Whether your orders, searches or recordings are among the proprietary data it trains on is the one thing neither document says, so the finding stands.
What this policy has done since we started watching
The record starts here. Amazon's 2 documents were first captured on 22 Aug 2026, and there's no earlier copy to compare them against, so this verdict has no history behind it yet. From now on every re-read is diffed against those copies, and anything that moves appears here.
Vendor response
We haven't yet written to Amazon 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