afnames

Afnames are anonymized name tokens used in data processing to replace real given names in datasets. They are designed to preserve some analytic utility—such as approximate name length, distribution, phonetic cues, or linkage possibility—while reducing the risk of revealing the actual names.

Origin and usage: The term appears in privacy-preserving data practices and synthetic data literature. Afnames may be created by deterministic pseudonymization, tokenization, or non-reversible hashing, possibly with a per-record salt or per-category salt to reduce re-identification risk. Some variants aim to maintain the ability to link records referring to the same individual across multiple datasets without exposing the real name.

Construction methods: Deterministic methods map a real name to a consistent afname within a given scope. Salted hashing or encryption with a secret key can produce non-deterministic afnames, improving privacy at the cost of cross-dataset linkage. Phonetic or distributional properties can be preserved to support name-based analyses, but such approximations can introduce biases.

Applications: Afnames are used in healthcare, social science research, and administrative data sharing where names are sensitive. They enable de-identified datasets for cohort studies, data linkages, and quality assurance, while limiting direct disclosure of personal identifiers.

Limitations: If the afname generation process is compromised or if the same seed is reused across datasets, re-identification risks increase. They are not a substitute for broader privacy protections or data minimization.

See also: Pseudonymization, anonymization, tokenization, k-anonymity.