Abstract:
The proliferation of Information and Communication Technologies (ICTs) exposes and recombines multiple facets of personal identity—biometric/passport, numeric cognitive, affiliative, functional social, and communi cative—through large-scale data collection, IoT sensor monitoring, online profiling, and cross-database link age. This paper offers a conceptual synthesis of how these processes erode anonymity and the right to limited access to the self (following Gavison), and formally reviews prevailing privacy frameworks—k anonymity, ℓ diversity, t closeness, and differential privacy. Using linkage attack reasoning and illustrative tabular exam ples, we show that record level anonymization provides only partial protection: multiplicity of identities does not preserve anonymity when AI enabled link ability can fuse disparate profiles; k anonymity is brittle under adversarial background knowledge; ℓ diversity and t closeness mitigate specific risks but often trade away utility and still fail in sparse distributions; and differential privacy bounds individual influence on statistical outputs yet does not directly prevent harms arising from indirect inference or real time transactional profiling. We argue for a system-level privacy by design paradigm that prioritizes data minimization, segmentation, and enforced unlikability; selectively applies differential privacy with policy-tuned ε; adopts multi-layer identity management with default pseudonymization; and institutionalizes transparency, meaningful consent, and con tinuous privacy risk monitoring. Reframing identity/anonymity protection from record redaction to ecosystem defenses against linkage and composition better aligns privacy safeguards with contemporary socio-technical infrastructures.