liveBook · Manning

Probabilistic identifier

As explained in Data for All

A probabilistic identifier is information that identifies a person or device with a probability greater than chance or a specified certainty threshold.

Could a few pieces of information identify you even if none names you outright? A probabilistic identifier is information that can identify a consumer or device with a degree of certainty “more probable than not.” A common interpretation is a greater than 50% chance of identifying one person. section 4.7.3

California’s CCPA includes probabilistic identifiers within its broad approach to personal information. That matters because combinations of data elements may point to an individual. The boundary remains disputed: the definition needs refinement before it can be applied consistently in practice. section 4.7.3

How Data for All covers it

How deep it goes
Defines the certainty threshold and notes unresolved practical boundaries, without giving an identification procedure.
Examples
Uses the possibility of identifying someone from a small number of data elements, without constructing a specific identifier.
What it assumes
Introduces the term through a plain-language certainty threshold rather than assuming a technical identification method.
What you can do after
Evaluate/understand why data that does not name someone outright may still fall within a privacy-law definition.
How it connects
Links the term to personal information, consumer privacy rights, the CCPA, and a proposed Massachusetts privacy law.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Data for All. The section links point to the parts of the book it draws on.