notes: research/papers/kerione-sustainable-privacy

This data as json

path title type status classification words body
research/papers/kerione-sustainable-privacy Sustainable Privacy summary v0.1 — 2026-10-14: drafted by Claude from the PDF (v1.2.8); unread by Mark; marked partner on Mark's word 2026-10-21, paraphrase only partner 476 # Sustainable Privacy **Abstract.** A 15-page white paper arguing that privacy over time needs economic, legal, regulatory and technical protections together, since K-anonymity de-identification and selective disclosure are defeated by re-identification attacks. It describes a contextual linkability re-identification attack on selective disclosure and proposes chain-link confidentiality and data loyalty, for readers concerned with privacy design and policy. ## Summary The paper defines sustainable privacy as the maintenance of personal data privacy over time, from two sides: how a person keeps their data private, and how a holder of someone else's data respects that privacy. Because any system that gathers, uses and shares information leaks, privacy erodes over time. Section 1 sets out the three-party exploitation model (first party source, second party intended recipient, third party anyone else), in which any unintended use of first-party data is exploitative, and refers to the SPAC (Secure Privacy, Authenticity, and Confidentiality) paper for third-party metadata correlation. It reviews privacy by design and by default and argues that K-anonymity de-identification is easy to defeat and should no longer count as best practice, while differential privacy and synthetic data are, so far, considered infeasible to defeat. Section 2 covers re-identification: linkage attacks and, as a new attack, contextual linkability. A verifier can shape the context of a presentation so that auxiliary data gathered at disclosure, combined with selectively disclosed attributes, re-identifies the discloser, defeating selective disclosure including zero-knowledge proof (ZKP) methods. The paper also discusses a security vulnerability of unlinkable disclosure. Section 3 covers protection of aggregated and non-aggregated first-party data and managing the time value of information. Section 4 turns to law and regulation: de-identification as a safe harbor is described as outdated; chain-link confidentiality attaches contractual "strings" to data for downstream users, and the ACDC (Authentic Chained Data Containers) standard is said to support it; data loyalty places a fiduciary-like duty on second parties; anti-assimilation rules would bar sharing and re-identification. It closes with Solove's view that privacy is about boundaries, not secrecy, and the "four Ds" of regulation. ## Key terms sustainable privacy, three-party exploitation model, K-anonymity, re-identification, contextual linkability, selective disclosure, chain-link confidentiality, data loyalty, anti-assimilation, time value of information ## Claims and their evidence - K-anonymity de-identification is easy to defeat and should no longer be considered best practice (section 1.2). [informative] - A contextual linkability attack can defeat selective disclosure, including disclosure using ZKPs (section 2.3). [informative] - The ACDC standard includes support for chain-link confidentiality (section 4). [informative] - Consent is not an effective mechanism for establishing a user's intent because it can be socially engineered (section 4). [informative] ## Related Secure Privacy Authenticity Confidentiality (SPAC); SPAC Presentation; the ACDC specification. ## Provenance Drafted by Claude on 2026-10-14 from the PDF (v1.2.8, 2026-08-11), read through the abstract, sections 1 and 1.1 to 1.2, 2.3, and section 4 to the end. Not yet read by Mark.

Links from other tables