The debate over government access to encrypted data has been framed as a security question — can exceptional access be provided without weakening the system? Lawful-surveillance protocols, an emerging cryptographic research area, aim to answer yes: government access without reduced systemic security. But Monteiro argues that even if these protocols succeed perfectly on their own terms, the problem is not solved. Security captures only part of what privacy protects.
Apple's discontinued client-side scanning initiative provides the clearest evidence. The technical objections were partly about security — could the system be subverted, expanded, misused? But the deeper resistance was about something security analysis cannot formalize: the transformation of a personal device into a site of pre-emptive inspection. The concern was not that the system would be breached but that it would work exactly as designed.
Reader note (2026-11-17): Checked against the full text of arXiv:2603.00841 (Monteiro, Security Is Not Enough, CSLAW 2026; v1), which I had earlier held as a weak read. The essay's account of the Apple discussion matches the paper's section 4.3, where the critics' report is quoted as saying that scanning all user devices without warrant or suspicion crosses a red line, and the paper reads that concern as going beyond security (it notes the report did not claim Apple's system was less secure than the server-side alternative). Two points differ in emphasis. The paper makes three arguments that security is not enough, not one: privacy as security does not explain the statutory conditions for interception (section 4.1), does not justify the positions opponents have taken on government hacking (section 4.2), and the client-side scanning objections (section 4.3); the essay presents Apple's system as the clearest evidence and leaves out the other two. And the paper describes its own contribution as modest: it says it aims only to show that there must be more to privacy than the security mold it has taken in encryption policy, and that it does not articulate here what an alternative conception of the right to privacy would be; the essay's phrase a negative result about the entire framing is a stronger description than the paper's own.
Privacy exceeds its security formalization the way meaning exceeds syntax. A protocol can be cryptographically sound, computationally secure, and formally verified while still violating something that resists technical specification. Monteiro's contribution is not a new protocol but a negative result about the entire framing: the security mold that privacy has taken in the encryption debate is structurally insufficient. Some properties of systems cannot be captured by the threat models those systems were designed to resist.
(arXiv:2603.00841)