How Topics helps to safeguard user privacy

Topics is designed to protect user privacy while enabling interest-based advertising. It reduces cross-site tracking and limits the sharing of sensitive data in several ways:

Reduce re-identification:

The Topics API makes it difficult to re-identify individuals across different websites. This is achieved through several mechanisms:

• Varied topic assignments: Different websites will usually receive different topics for the same user in the same week. This makes it harder to link a user's activity across multiple sites.
• Weekly updates: Topics are updated weekly, limiting the rate at which information can be gathered about a user's interests.
• Randomization: A fraction of the time, a random topic will be returned for a given site, adding another layer of protection against re-identification.

Reduce sharing of sensitive data:

The topics revealed by the Topics API are significantly less personally sensitive than what could be transmitted using cookies or covert tracking methods.

• Human-curated taxonomy: The taxonomy of topics is human-readable and human-curated and excludes sensitive categories like health, race, and sexuality.
• Ongoing discussion: The list of topics is publicly shared so that users and the community can evaluate whether they represent a risk to sensitive categories.