Amazon's privacy notice now says it uses data to develop generative AI

The change

The Amazon.com Privacy Notice carries a new date: "Last updated: June 30, 2026." The version it replaces was dated December 31, 2025, and Amazon still links to it as the prior version. The notice covers Amazon's stores, devices, apps, and services such as Prime Video and Alexa.

The edits are few. Compared with the December version, the notice gains one passage about artificial intelligence, two lines about advertising data, and one line about linked accounts, apart from small wording fixes. The December version did not contain the words artificial intelligence, machine learning, or generative AI at all.

What changed

• Generative AI. Under the purpose of providing and improving its services, Amazon now says it uses machine learning and artificial intelligence "to support and enhance the customer experience," and adds: "This includes using data to develop or deploy our generative AI models and services."

• A new disclosure. The same passage links to a separate page, the Generative AI Development Disclosure, which is also added to the list of related documents at the end of the notice.

• Ad interactions elsewhere. Among the information Amazon receives from other sources, the old line "search results and links, including paid listings (such as Sponsored Links)" was replaced by information about how you interact with "ads served by Amazon on third-party websites and apps."

• Third-party ad data. A new line covers content received from third parties, "such as advertising-related information and identifiers."

• Linked accounts. The list of moments when you give Amazon information now includes when you "use Login with Amazon or otherwise link your Amazon account to a third-party business."

A short passage inside an existing purpose

The way the AI passage was added matters as much as what it says. Amazon did not create a new purpose with its own heading. It extended a bullet that was already there, the one about using personal information to "improve the usability and effectiveness of the Amazon Services," and opened the addition with the words "In so doing."

Read that way, developing generative AI is presented as part of improving the services, not as a separate use that needs a separate permission. The notice still opens with the statement that by using Amazon Services "you are consenting to the practices described in this Privacy Notice."

The new sentence speaks of "using data" and does not say which data. For that, it sends the reader to the disclosure. AI training language has been appearing in other large vendors' documents too, as when Apple added an exception to its promise not to train AI on your data.

What the Generative AI Development Disclosure says

At the time of publication, the disclosure describes the data Amazon and its affiliates use to develop or deploy generative AI. It says Amazon may train and test on licensed and proprietary datasets, synthetic datasets, open-source datasets, and publicly available content, including web-crawled data. It says the services may run on foundation models built by Amazon or by third parties.

Two statements stand out for a privacy reader. The first is about timing: "We have been collecting data since before 2022." The second is that Amazon describes safeguards meant to limit the impact of any processing of personal information in training, which means the disclosure does not rule such processing out.

This disclosure sits on Amazon's retail help pages. Amazon Web Services publishes its own disclosure for its cloud products, and the two are separate documents.

The advertising lines

The advertising edits are easier to miss because they sit in the long list of examples at the bottom of the notice. Before, the entry in that spot was about search results and paid listings. Now there are two entries: one for how you respond to Amazon's ads on sites and apps that are not Amazon's, and one for advertising information and identifiers that third parties send to Amazon.

The notice says Amazon receives this content "to provide, develop, and improve Amazon Services." Other large vendors have widened their ad-data language this year as well, including when Uber said its ad partners may combine your data with other sources and when OpenAI widened what it does with data from advertisers.

What did not change

• The section "What Choices Do I Have?" is word for word the same, and it does not mention AI.

• Amazon still says: "we are not in the business of selling our customers' personal information to others."

• The notice still says: "We do not use information that personally identifies you to display interest-based ads."

• The closing section still promises that Amazon "will never materially change our policies and practices to make them less protective" of information collected in the past "without the consent of affected customers."

Why this matters

Amazon is a vendor to a great many companies in some form, including companies that would never list it as one. Employees order supplies from it, sign in to other tools with Login with Amazon, and use its devices and apps on the same phones and laptops they work on. The privacy notice is the document that governs all of that.

For a SaaS company, the update is a clear example of how AI language now enters vendor documents. There was no new agreement and no new heading. A short passage was attached to a purpose that customers had already accepted, and the detail was placed on a page that can change on its own schedule. A review that only looks for new sections would not have found it.

That is why AI clauses belong in continuous monitoring of vendor legal changes and not in a one-time vendor review. The next edit may be to the disclosure page, which the notice now relies on.

Potential impact

For a company whose people use Amazon in their work, the update raises five practical questions:

• Which company or shared Amazon accounts exist, and which are linked to other services through Login with Amazon?

• Do any Amazon devices or apps in the workplace handle voice, images, or documents that the company would not want used for model development?

• Has anyone read the Generative AI Development Disclosure, and is someone watching it for changes?

• Does your own vendor inventory treat Amazon's consumer services and AWS as separate vendors with separate documents?

• When a vendor adds AI language inside an existing clause, would your current review process notice?

How Venpo detected it

Venpo monitors Amazon's privacy notice as part of continuous vendor risk monitoring. It marked the added AI passage, the replaced advertising line, and the two new entries against the December 31, 2025 version, with the rest of the notice unchanged. Every quote in this article was checked against the redline and the live page.

Business outcome

Teams that track Amazon saw exactly which sentences were added and where. That turns a long notice into four edits a person can read in a minute, and it shows that the choices section was left alone. Without the comparison, the visible signs of the update are a new date at the top and a link to the prior version.

Key takeaway

Amazon's privacy notice now says it uses data to develop or deploy generative AI, lists more advertising data from outside Amazon, and names linked accounts as a source of information. The AI passage sits inside a purpose customers had already accepted, and the way to see edits like that is to monitor vendor terms and policies as a routine.