Inferential Disclosure Through Joined Fragments: How Separate Pieces Can Say What None of Them Said

This conceptual paper introduces Inferential Disclosure Through Joined Fragments, a mechanism by which separately disclosed pieces of information become available to one reader, who can connect them to infer a relationship that no piece established alone. The paper asks how strongly the evidence supports the connection and whether the purposes for which the fragments were disclosed permit their combined use. A reader may have strong grounds for an identification while permission to make that use of the records remains unsettled. The first case concerns the Japanese designator “AF” before the Battle of Midway. By March 1942, analysts at Corregidor had identified AF as Midway, while disagreement persisted in Washington. In May, Midway sent an uncoded message reporting a shortage of fresh water. A subsequent Japanese report attributed a water shortage to AF. The shared detail connected the two messages and corroborated the earlier identification. The second case concerns AOL’s 2006 release of search queries grouped under user numbers. One number brought together searches mentioning a town, a surname, and a housing subdivision. Reporters followed those clues to Thelma Arnold, who recognized the searches as her own. She explained that she had researched some symptoms for friends. The evidence supported identifying who had conducted the searches while leaving open whom any particular search described. The paper places this case alongside Latanya Sweeney’s work on re-identification through linked records. Part of the Sources and Evidentiary Limits series within On Human Modeling, the paper also offers a distinction relevant to human–AI interaction. A draft exhibition program might list a title and booking code, while a venue’s schedule lists the same code and a room. An AI assistant could infer where the exhibition is planned. Whether the organizers authorized combining those records for a public announcement is a separate question.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-18
DOI
https://doi.org/10.5281/zenodo.24355328
Primary Topic
Privacy, Security, and Data Protection
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Inferential Disclosure Through Joined Fragments: How Separate Pieces Can Say What None of Them Said

Alexandra Cary
2 links
Zenodo (CERN European Organization for Nuclear Research)
Privacy, Security, and Data Protection
article

Inferential Disclosure Through Joined Fragments: How Separate Pieces Can Say What None of Them Said

Alexandra Cary
article en

Abstract

This conceptual paper introduces Inferential Disclosure Through Joined Fragments, a mechanism by which separately disclosed pieces of information become available to one reader, who can connect them to infer a relationship that no piece established alone. The paper asks how strongly the evidence supports the connection and whether the purposes for which the fragments were disclosed permit their combined use. A reader may have strong grounds for an identification while permission to make that use of the records remains unsettled. The first case concerns the Japanese designator “AF” before the Battle of Midway. By March 1942, analysts at Corregidor had identified AF as Midway, while disagreement persisted in Washington. In May, Midway sent an uncoded message reporting a shortage of fresh water. A subsequent Japanese report attributed a water shortage to AF. The shared detail connected the two messages and corroborated the earlier identification. The second case concerns AOL’s 2006 release of search queries grouped under user numbers. One number brought together searches mentioning a town, a surname, and a housing subdivision. Reporters followed those clues to Thelma Arnold, who recognized the searches as her own. She explained that she had researched some symptoms for friends. The evidence supported identifying who had conducted the searches while leaving open whom any particular search described. The paper places this case alongside Latanya Sweeney’s work on re-identification through linked records. Part of the Sources and Evidentiary Limits series within On Human Modeling, the paper also offers a distinction relevant to human–AI interaction. A draft exhibition program might list a title and booking code, while a venue’s schedule lists the same code and a room. An AI assistant could infer where the exhibition is planned. Whether the organizers authorized combining those records for a public announcement is a separate question.

Zenodo (CERN European Organization for Nuclear Research)
Privacy, Security, and Data Protection
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon