What Amazon clarified on 9 October

On 9 October, Amazon told the media that the details shown in Amazon About You are inferred only from a customer’s activity within Amazon: purchases, searches, saved Lists, reviews, and conversations with its shopping assistant. The company specifically denies using outside data to construct this profile; that claim has not been independently audited. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026]

The feature itself is not new. Amazon introduced it on 13 May as a single place to view and edit information that shapes recommendations. Its announcement listed purchase history, searches, Lists, reviews, and shopping-assistant conversations, and said customers could update or exclude individual details from personalization. [2 · Amazon · official About You announcement, 13 May 2026]

The new clarification draws an important boundary: removing a detail from Amazon About You means it should no longer be used for recommendations, but it does not erase the purchase, search, or other activity from which the inference arose. Amazon’s Privacy Notice describes broader collection, including automatic service-use information and data from other sources; the 9 October statement concerns the About You profile, not the company’s entire data infrastructure. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice]

Control over a recommendation is not data deletion

For Amazon’s marketing, Amazon About You turns hidden assumptions in a recommendation system into an editable part of the customer relationship. A shopper can correct an outdated or wrong inference, while the platform receives a direct preference signal. Public materials do not explain how quickly an edit changes recommendations, whether it applies across every shopping surface, or whether a customer can inspect the reason for each inference. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026]

There are three distinct levels of control: correct the displayed detail, stop using it for personalization, and delete the underlying history. Amazon confirms the first two actions for About You but does not promise the third in the same interface. Its Privacy Notice places data deletion and advertising controls in separate mechanisms, so editing the profile should not be described as erasing the customer data trail. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice]

Amazon has not published the feature’s usage rate, share of corrected inferences, change in recommendation accuracy, conversion, repeat purchase, churn, or customer lifetime value (LTV). There is no observation period, sample, comparison baseline, or control group. The renewed attention therefore reveals interest and a trust risk, but proves neither commercial success nor damage to sales. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice]

How to measure trust rather than curiosity

Viral attention may temporarily increase visits to a settings page, but that is not the same as a better customer relationship. More useful measures are the share of customers who inspect the profile, disagreement frequency, removal rate, speed of recommendation changes after an edit, and subsequent use of personalization. Those measures should be evaluated alongside complaints, personalization opt-outs, and long-term purchases. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice]

A 491-participant experiment published in the Journal of Computer-Mediated Communication found that meaningful explanations of AI decisions reduced uncertainty and increased trust; placebic explanations behaved differently depending on the system type. This was a laboratory result in other tasks. It supports the mechanism that a useful explanation can help, but it does not establish an Amazon About You effect on conversion or retention. [4 · Journal of Computer-Mediated Communication · AI transparency experiment, 2021]

A causal test could randomly assign comparable customers to standard recommendations, an editable profile, or an editable profile that explains the origin of each inference. Primary outcomes should not be settings-page clicks but feedback-based recommendation accuracy, incremental gross profit, repeat purchase, and complaints. Shared household accounts need separate analysis because purchases by different people can produce plausible but wrong attributes. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice] [4 · Journal of Computer-Mediated Communication · AI transparency experiment, 2021]

Expert commentary

Amazon made part of its personalization visible just as that visibility became a reputational event. This is a useful shift: customers gain a place to object, and the company can correct its model with more than click behavior. Because the clarification followed mass attention, however, it currently looks more like trust management than evidence of a mature feedback system. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026]

The decisive boundary is between an inference and the data behind it. Removing a sentence from the profile is not the same as deleting the history that produced it. If the interface does not explain this beside the action, customers may expect more control than they receive. Durable relationships require distinct, understandable commands to correct a preference, stop its use, and delete source records where retention obligations allow. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice]

Research helps explain why merely displaying a profile is insufficient. In a 491-participant experiment, trust increased when an explanation was meaningful and reduced uncertainty. The effect size cannot be transferred to Amazon because the tasks, interface, and consequences differ. The modest practical lesson is that an explanation should answer “why did the system decide this about me?” rather than merely restating that personalization occurs. [4 · Journal of Computer-Mediated Communication · AI transparency experiment, 2021]

The business value will emerge only if corrections improve the next choice. Amazon should compare recommendations before and after an edit against a held-out control group, measuring incremental conversion and gross profit and then repeat purchase. Page views and edited details are behavioral signals, not proof of economic return and certainly not direct measures of trust. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026]

Risk is especially visible in shared accounts, gift purchases, and major life changes, where behavioral history maps poorly to one person. A wrong attribute may be harmless, or it may touch health, body, or family circumstances. The more sensitive the inference, the stronger the need for data minimization, an explanation, rapid correction, and suspension of the disputed attribute until it is checked. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice]

Over the next 6–12 months, convincing evidence would be operational discipline rather than another feature presentation: the share of explained inferences, time for an edit to take effect, recurrence of errors, personalization opt-outs, complaints, recommendation quality, and retention after correction. Until those data exist, Amazon About You is an interesting transparency mechanism with limited control, while its effect on trust and commerce remains a testable hypothesis. [1 · Business Insider · Amazon clarification on About You data, 9 October 2026] [2 · Amazon · official About You announcement, 13 May 2026] [3 · Amazon · current Privacy Notice] [4 · Journal of Computer-Mediated Communication · AI transparency experiment, 2021]

Sources

  1. Business Insider · Amazon clarification on About You data, 9 October 2026 — The fresh event source: reports Amazon’s position on where the inferences come from, editing details, and retention of the underlying history. The outlet did not audit Amazon’s systems.
  2. Amazon · official About You announcement, 13 May 2026 — Primary source on the feature’s purpose, personalization inputs, and customer controls. It is the company’s description, not an independent outcome assessment.
  3. Amazon · current Privacy Notice — Primary source on broader collection and use of data across Amazon services, personalization, advertising, and account controls. It shows that the About You clarification does not describe Amazon’s entire data environment.
  4. Journal of Computer-Mediated Communication · AI transparency experiment, 2021 — A 2×3×2 experiment with 491 participants found that meaningful explanations reduced uncertainty and increased trust in a decision system. It did not study Amazon, e-commerce, or actual retention.