P3(b). Inference of user information

The company should clearly disclose what user information it infers and how.

Elements:

1. Does the company clearly disclose all the types of user information it infers on the basis of collected user information?
2. For each type of user information the company infers, does the company clearly disclose how it infers that user information?
3. Does the company clearly disclose that it limits inference of user information to what is directly relevant and necessary to accomplish the purpose of its service?

Clearly disclose(s) — The company presents or explains its policies or practices in its public-facing materials in a way that is easy for users to find and understand.

Collected user information — User information that a company either observes directly or acquires from a third party.

Data inference — Companies are able to draw inferences and predictions about the behaviors, preferences, and private lives of its users by applying “big data” analytics and algorithmic decision making technologies. These methods might be used to make inferences about user preferences or attributes (e.g., race, gender, sexual orientation), and opinions (e.g., political stances), or to predict behaviors (e.g., to serve advertisements). Without sufficient transparency and user control over data inference, privacy-invasive and non-verifiable inferences cannot be predicted, understood, or refuted by users. For more see: Wachter, Sandra and Mittelstadt, Brent, A Right to Reasonable Inferences: Re-Thinking Data Protection Law in the Age of Big Data and AI (October 5, 2018). Columbia Business Law Review, 2019(2), https://ssrn.com/abstract=3248829

User information — Any data that is connected to an identifiable person, or may be connected to such a person by combining datasets or utilizing data-mining techniques. User information may be either collected or inferred. As further explanation, user information is any data that documents a user’s characteristics and/or activities. This information may or may not be tied to a specific user account. This information includes, but is not limited to, personal correspondence, user-generated content, account preferences and settings, log and access data, data about a user’s activities or preferences collected from third parties either through behavioral tracking or purchasing of data, and all forms of metadata. User information is never considered anonymous except when included solely as a basis to generate global measures (e.g. number of active monthly users). For example, the statement, ‘Our service has 1 million monthly active users,’ contains anonymous data, since it does not give enough information to know who those 1 million users are.

Indicator guidance: In addition to collecting information about users, companies also perform big data analytics to infer additional data points on the basis of the collected information. This inferred information is then used for a variety of purposes, much in the same way as collected information. In addition to disclosing the information that they collect, disclosing the purpose for which they collect it, and committing to only collect information that is relevant and necessary to provide their service, companies should also disclose what information they infer and how they infer it. They should also commit to only infer information that is relevant and necessary to provide the service. For example, companies should not try to infer their users’ religion, sexual orientation, or health status (such as by assigning them to an audience category based on this characteristic) unless that information is somehow directly necessary to accomplish the purpose of their service.

• Company privacy policy, cookies policy
• Company web page or section on data protection or data collection