What empirical studies measure how encryption adoption affects the number of CSAM reports and law enforcement case openings?
Executive summary
There are remarkably few rigorous empirical studies that directly measure how adoption of end-to-end encryption (E2EE) changes the number of CSAM reports or the rate at which law enforcement opens cases; most available work is descriptive, agency reporting, or policy analysis pointing to gaps in evidence [1] [2] [3]. Where researchers have examined outcomes, the results are mixed or inconclusive, and numerous expert and government reports explicitly say data are inadequate to draw causal conclusions [1] [3] [2].
1. What the peer‑reviewed empirical literature actually measures
A notable empirical contribution is the Dutch study published in Crime Science (March 2023), which examined the impact of E2EE on criminal-court case outcomes and concluded that the “going dark” hypothesis—E2EE catastrophically reducing prosecutorial outcomes—lacks empirical support in that context; the paper emphasizes law enforcement still uses alternative investigative avenues such as cell‑tower and metadata evidence [1]. Beyond that study, peer‑reviewed literature directly quantifying changes in CSAM reports or case openings after encryption rollouts is essentially absent in the sources provided; most academic pieces document mechanisms and legal trade‑offs rather than producing causal estimates [1] [4].
2. Industry and government reports: lots of counts, little causality
Electronic service providers and clearinghouses publish large counts of CSAM reports and removals, and government or industry analyses document dramatic volumes—NCMEC numbers and platform takedowns are cited repeatedly—but those publications rarely isolate encryption as the causal driver of changes in reporting or case initiation [2] [5]. Australian Institute of Criminology reporting, for example, notes huge numbers of proactive detections and removed files while also warning that empirical evaluations of detection technologies are scarce, underscoring that quantity metrics do not demonstrate how encryption adoption alters those flows [2] [5].
3. Policy and think‑tank assessments that summarize evidence gaps
High‑level policy reports from CSIS and the Carnegie Endowment conclude that while encryption complicates some investigative techniques, the public‑safety risk “has not reached a level” warranting design mandates and that available data are insufficient to evaluate many proposed fixes; both call for better data collection, analytics, and capacity building rather than reflexive mandates to weaken encryption [6] [3]. These assessments frame the empirical deficit as a central problem in the debate: stakeholders know counts but not counterfactuals [3].
4. Law enforcement and advocacy perspectives that interpret limited evidence differently
Law‑enforcement and some advocacy voices assert stronger effects: U.S. DOJ and security‑oriented groups warn that encryption enables offenders to hide communications and frustrates investigations [7] [8], and industry‑funded or policing stakeholders argue E2EE will “drastically reduce” the amount of CSAM reported to authorities [9]. These claims are empirical in tone but, according to the literature and reviews cited, are often based on operational experience and logic rather than published causal studies that measure pre/post adoption effects [1] [3].
5. Fragmentary empirical signals and the limits of current evidence
A few analytic claims—such as an estimate that Facebook could not identify a large share of CSAM after E2EE changes—appear in policy briefs and case studies (e.g., a cited Koomen figure of ~70%), but those are not equivalent to peer‑reviewed causal estimates linking platform encryption rollouts to national CSAM reporting or case‑opening trends [4]. Multiple sources explicitly conclude empirical evaluations of detection technologies and encryption effects are scarce, meaning that existing counts and expert testimony cannot substitute for studies that compare reporting and investigative outcomes before and after encryption while controlling for confounders [2] [5] [3].
6. Bottom line and where research should go next
The bottom line is that empirical evidence measuring the causal effect of encryption adoption on CSAM report volumes and law enforcement case openings is thin: a few descriptive and country‑level studies (e.g., the Dutch court‑outcome analysis) exist, but no comprehensive, multi‑jurisdictional causal studies are cited in the available material [1] [3]. The consistent recommendation across policy reviews is straightforward—collect standardized data on encrypted devices encountered, transparency from platforms about detection-before‑and‑after encryption changes, and funded, independent evaluations—so researchers can move from plausible assertions to measurable effects [3] [6].