Navy SEALs and the War Department's Signal Intelligence Photo Vault
By Marcus Boone ·
Masked as a public relations archive and kept under the stewardship of a signal intelligence division, these military images served as a secret baseline for targeting.
A man in a window in Baghdad does not know he has become a data point until the missile arrives. He is not a person to the analyst; he is a pattern of movement, a shape in a frame, a confirmation of a coordinate. The cost of this translation from human to target is the total erasure of privacy in the name of operational security.
I am working from the archival scope-and-content description of these records, a catalogue of what survives rather than the photographs themselves. The record identifies the collection as the "Combined Military Service Public Affairs Photographs." On its face, this is a scrapbook of military life—retirement ceremonies, museum exhibits, and the mundane activities of the U.S. Army, Navy, Air Force, Marine Corps and Coast Guard. But the record contains a detail that transforms this benign collection into something far more predatory. These records were not held by a press office or a historical society. They were "originally maintained by the Signal Intelligence Service Division under the Office of the Chief Signal Officer within the War Department," the administrative body that managed the nation's military forces prior to 1947.
The Navy SEAL Project Book
Public affairs offices produce brochures; signal intelligence divisions produce targets. When a record lists "U.S. Army Portraits" alongside the "Navy SEAL Project Book," the analyst sees a discrepancy in purpose. The Navy SEALs, the elite maritime special operations force specializing in unconventional warfare, do not typically lend their operational documentation to a public relations effort. The presence of the "Navy SEAL Project Book" and "Combat Camera" assets within a SIGINT vault suggests that the "Public Affairs" label is a skin, not a substance.
This is a collection of visual intelligence. The record lists specific sites: "Port Hueneme," a primary hub for the Army's logistics and port operations; a "Ship repair facility in Guam," a critical strategic node in the Pacific; and "Vandenberg Air Force Base missile launches." These are not the subjects of postcards. They are the structural markers of American power and the blueprints of the environments where the military operates. By housing these images within the Signal Intelligence Service Division—the unit responsible for intercepting and analyzing electronic signals—the War Department ensured that the visual "ground truth" was married to the electronic data flowing into the desk.
The Signal Intelligence Vault
The administrative handling of the file reveals a preoccupation with categorization that exceeds the needs of a historian. The record notes that the Still Depository Accession Number (SDAN) assigned by DIMOC, the administrative body responsible for cataloging these assets, uses a specific coding system: "C" for color negative, "T" for transparency, "N" for black-and-white negative, and "P" for print. This level of technical granularity is the hallmark of a database designed for rapid retrieval and cross-referencing, not a museum archive.
Consider the connections. The record links the "Cobra Gold '95 exercises" and "Exercise Freedom Banner" to the same custodial chain as the "9/11 Attack Collection." The pattern here is the integration of disparate data streams. Cobra Gold is a massive multilateral exercise in Thailand; Freedom Banner is a strategic maneuver in Korea. To a public affairs officer, these are opportunities for a photo op. To the Signal Intelligence Service Division, these are opportunities to map the physical layout of allied installations and the movement of assets in contested theaters.
By centralizing imagery from all five military branches under a single SIGINT authority, the War Department created a master visual index. The desk's reading is that the "Public Affairs" designation was a classification workaround. It allowed the military to distribute these images internally across various commands without triggering the suffocating security protocols of a formal intelligence file. It gave the imagery the mobility of a press release while maintaining the control of a spy agency.
The Maven Blueprint
This analog vault is the direct ancestor of the modern machine. The public record establishes that Project Maven, launched in 2017, is a Department of Defense initiative designed to accelerate machine learning in intelligence workflows, specifically using computer vision to flag targets from drone and satellite imagery. Maven does not invent the concept of visual targeting; it simply automates the process that the Signal Intelligence Service Division was practicing with film and transparency.
If this file is shaped the way it looks, the "Combined Military Service Public Affairs Photographs" was the manual version of a training dataset. For a machine learning model to identify a "military asset," it first needs thousands of labeled examples of what that asset looks like in the real world. The "Navy Aerials," the "Ship repair facility in Guam," and the "V-22 Flight Test" photos provided those labels. They provided the baseline against which intercepted, low-resolution imagery could be verified.
The irony is thick. The records list the "Vietnam War POW Return" and "Prisoners of War/Missing in Action (POW-MIA) Accounting"—images of the military's most profound failures and heartbreaks. Yet these images sat in the same vault as the tools used to facilitate the next generation of strikes. The human cost of war is archived in the same folder as the technical means of delivering it.
The Pattern of the Gap
What is missing from the record is as telling as what remains. We have the list of photographs, but we do not have the cross-referenced intelligence reports. We have the "Navy SEAL Project Book," but not the operational notes that explain why those specific images were valuable to a signal officer. The gap is shaped exactly like a target folder. The public affairs photographs were the visible half of a ledger; the other half—the metadata, the target coordinates, and the operational justifications—remains classified or was destroyed.
The pattern suggests that the War Department built a VISINT database masquerading as a PR archive to avoid the friction of the bureaucracy it served. By labeling the work as "Public Affairs," they could move visual assets between the Army, Navy, and Air Force with ease, while the Signal Intelligence Service Division held the master key. They turned the military's public image into a weapon of reconnaissance.
The desk's reading is that this archive proves the Department of Defense has always viewed "public" information as a primary source for clandestine targeting. A full release of the original metadata would likely show that these "public affairs" photos were used to calibrate the sensors and the analysts who decided who lived and who died in the decades that followed. The redacted pages are not protecting the privacy of the soldiers in the photos; they are protecting the methodology of the hunt. The blueprint was drawn in the War Department's signal vault, and Project Maven is simply the digital realization of a long-standing appetite for total visual dominance.