Security Is Not Enough: Privacy in Encryption Regulation and Lawful-Surveillance Protocols

AI summary

Overview

Research area: Computer science and law — specifically encryption regulation, lawful-surveillance protocols, and the normative (not technical) content of privacy in encryption policy. Listed under cs.CR on arXiv.

Technical level: Intermediate. The cryptography discussed is described conceptually (key escrow, LEAF, secure multiparty computation, private information retrieval), but the argument is legal, historical, and normative, so readers need some familiarity with the encryption-policy debate rather than mathematical cryptography.

Scope (one sentence): The paper argues that the dominant "privacy-as-security" framing in encryption policy is insufficient to evaluate lawful-surveillance protocols or to explain existing positions on exceptional access, and that a richer account of privacy is therefore required.

What This Paper Is About

Governments want lawful access to encrypted data, and opponents of that access have increasingly reframed the fight from "privacy versus security" to "security versus security" — arguing that backdoors and exceptional-access mechanisms would harm everyone's security, including law enforcement's. A research agenda called lawful-surveillance protocols tries to use cryptography to give officials access without damaging systemic security.

This paper's core problem is that the security framing, however strategically effective, cannot fully capture what is at stake. The author argues that a richer conception of privacy is needed both to assess policy and to guide research on lawful-surveillance protocols, and that objections to Apple's abandoned client-side scanning system already point beyond security.

Key Contributions

1. Argues privacy-as-security cannot evaluate lawful-surveillance protocols. The paper contends that a conception of privacy as security has little to say about the cryptography research agenda that aims to enable government exceptional access without compromising systemic security — and that this limitation is not contingent on whether that agenda succeeds.

2. Shows the current normative landscape is inexplicable under a security-only account. The paper argues that privacy-as-security does not explain why exceptional-access mechanisms could not offset introduced risks by modifying background institutional conditions, such as statutory requirements for interception (Section 4.1), and does not justify the exceptional-access positions opponents have taken on government hacking (Section 4.2).

3. Uses client-side scanning as a test case. The paper argues that fundamental objections by long-time critics of exceptional access to Apple's proposed and abandoned client-side scanning system themselves require an understanding of privacy that goes beyond security (Section 4.3).

4. Provides a historical and survey base. It recapitulates the Crypto Wars (Clipper Chip, "Going Dark" and "Apple v. FBI," and a possible third chapter) and surveys lawful-surveillance protocol proposals, concluding that "a full articulation of privacy was never really offered." The author explicitly states the contribution is modest and that the paper does not articulate what an alternative conception of the right to privacy would be.

Main Findings

• The opposition reframed the dispute as security versus security. Rather than privacy versus security, opponents argue exceptional access should be rejected because it would undermine the very mission of law enforcement, compromising citizens' and national security. The paper credits Susan Landau with developing this reframing, appearing in her 2016 testimony before the United States House of Representatives and in her 2017 book, where she writes: "The government's role is to provide security—national security and law enforcement—and not to prevent individuals from maintaining their own security."

• The reframing removes law enforcement's "home-field advantage." Drawing on Solove, the paper describes privacy as playing "away games" on fields designed around security, where privacy advocates also lack information due to secrecy in national security and public safety matters.

• Lawful-surveillance protocols fit the security mold and are anchored in procedural safeguards. Proponents describe combining cryptographic protocols (SMC, PIR) with blackletter law, aiming for "lawful, accountable, privacy-preserving surveillance," and draw on Solove's idea of reconciling privacy and security through oversight, limiting future uses of personal data, and controlled implementation. The author notes privacy-as-security "has been the only articulation to gain traction."

• Courts have largely embraced a pragmatic, procedure-focused approach. In Podchasov v. Russia (European Court of Human Rights, Third Section, 2024), the court held that providers could not be required to build systems to maintain decryption capabilities, seizing on impacts on the safety of all users to find the Russian requirement disproportionate. A commentator noted the judgment "seems to leave options open if a decryption technology becomes available that would not weaken the security of all users."

• The Clipper Chip was not brought down by a security flaw exposing encrypted data. The National Academies report cited states such a flaw was never found. A vulnerability that let users defeat the escrow itself by spoofing the LEAF made front-page news in the June 2, 1994 New York Times, but strong opposition coalesced around costs to technological innovation and a nascent digital economy, including a Computer Professionals for Social Responsibility petition signed by more than 50,000 people, and concerns that escrowed encryption would hurt U.S. global competitiveness under export controls that limited exports to 40-bit keys.

• Historical details of the first Crypto Wars. The Clipper Chip implemented the NSA-designed Skipjack algorithm for escrowed encrypted telephony; the Capstone Chip implemented Skipjack for computer communications. The Escrowed Encryption Standard encrypted 80-bit session keys, with the Law Enforcement Access Field (LEAF) carrying the session key and a chip unique identifier, encrypted with a family key common to all chips plus a chip unique key; regulations designated NIST and a division of the Treasury Department as escrow agents. Clipper was the NSA's reaction to AT&T's planned $1,295 Telephone Security Device Model 3600 (TSD-3600). The Department of Justice alone acquired 9,000 TSD-3600E devices at about $1,000 per unit, while the Clipper Chip itself cost between $15 and $25 per unit. International efforts through the 1996 Wassenaar Arrangement (then limiting exports to 64-bit symmetric keys, loosened in 2009) and the OECD (backed only by France and the U.K.) failed.

• Deployment lag and the second Crypto Wars. In 2016 only 40% of internet traffic was encrypted with Transport Layer Security; the figure would reach 80% in 2019. In late 2014, after the Snowden revelations, Apple, Google, and Meta's WhatsApp announced end-to-end encryption by default. FBI General Counsel Valerie Caproni's 2011 "going dark" warning was echoed by Director James Comey at the Brookings Institution a month after Apple's announcement. In Brazil, courts ordered WhatsApp suspended in 2015 after it deployed E2EE; the Supreme Court stayed the order, and constitutional challenge ADPF 403 remains pending. The San Bernardino attack killed 14 victims.

• A concrete motivation for protocol design. In the "high country bandits" case, a 2010 bank robbery spree across three cities in Arizona and Colorado led the FBI to compare cell-tower connection lists, disclosing the metadata of 149,999 people. The proposed alternative would let the carrier keep connection information encrypted while returning only suspects recurring across datasets, with the FBI's query itself encrypted so the carrier would not know the target.

• Protocol designs surveyed so far. Stefen Savage's self-escrow proposal lists four properties — non-scalability, authorization, particularity, and transparency — requires physical possession of the device for a period before the mechanism operates, self-escrows the key on the device, displays a message upon boot if an unlock is attempted, and relies on manufacturer cooperation as an external check. Ray Ozzie's "CLEAR" proposal, discussed in confidential informal conversations and publicized in a 2018 Wired article by Steven Levy before Ozzie posted related slides, is also described as based on self-escrow.

• No empirical or quantitative evaluation is reported. The provided paper content is a normative and doctrinal argument; it reports no benchmarks, datasets, or performance measurements. The supplied text also ends mid-discussion of Ozzie's CLEAR proposal, so the detailed contents of Sections 3.3 through 4.3 are not reported in the content available here, beyond the roadmap statements in the introduction.

Methodology in Plain English

This is not an experimental paper. The researcher works the way legal and policy scholarship does: by reconstructing history, reading documents, and analyzing arguments.

• Historical reconstruction: The paper traces encryption regulation from the 1970s (including an NSA letter warning that a 1977 IEEE information theory symposium would violate the International Traffic in Arms Regulations, a temporary NSF suspension of new cryptography funding, and deletion requests over Kahn's 1967 Codebreakers) through the Clipper Chip, the "going dark" period, and current disputes.
• Document and literature survey: It reads congressional testimony, government reports (including the congressionally commissioned National Research Council report Cryptography's Role in Securing the Information Society, known as CRISIS), civil society petitions, court filings, and the academic canon on encryption policy.
• Comparative case law analysis: It examines European Court of Human Rights and Court of Justice of the European Union decisions alongside U.S. disputes.
• Argument analysis of proposals: It organizes lawful-surveillance protocol proposals into categories (data already available in cleartext; data stored in encrypted devices; self-escrow designs) and examines what conception of privacy each presupposes.
• Conceptual argument: The central method is showing that a given framing (privacy-as-security) fails to explain positions that its proponents actually hold, which the author treats as evidence that the framing is incomplete.

Why This Matters

Impact on research. The paper puts a normative burden on a technical agenda: if lawful-surveillance protocols are justified by security alone, their designers and advocates are left without an account of what else is at stake. It also pushes privacy scholarship to move past procedural and security-based accounts, arguing that "there must be more to privacy than the security mold it has taken."

Real-world applications.

• Ongoing encryption regulation battles: the EU's Child Sexual Abuse regulation proposal (COM(2022) 209 final, informally

Authors’ abstract

This article argues that security is not enough to fully capture what is at stake in government exceptional access to encrypted data. A conception of privacy as security has little to say about ``lawful-surveillance protocols'' -- an active research agenda in cryptography that aims to enable government exceptional access without compromising systemic security. But the limitations are not contingent on the success of this agenda. The normative landscape today cannot be explained if security is all there is to privacy. And fundamental objections to Apple's abandoned client-side scanning system gesture beyond security. This article's contribution is modest: to show that there must be more to privacy than the security mold it has taken. A richer understanding is needed both to assess policy and to guide research on lawful-surveillance protocols.