Three steel filing cabinets sit in the basement of the Chernihiv Regional State Administration. They hold roughly 14,000 digital images on unmarked USB drives. Liudmyla Kovalenko carried those drives downstairs in March 2022, wrapped in a wool blanket, while Russian artillery struck the building’s upper floors. She headed the municipal archive’s digitization unit, but she had no formal training in conflict documentation. She was a records manager who, in the invasion’s first week, started walking to every damaged structure she could reach and photographing it. Her images are blurry, uncomposed, shot on a phone with a cracked lens. They now form the most complete ground-level record of destruction in the city’s central district. Prosecutors have used them in three war crimes cases. The World Bank’s Rapid Damage and Needs Assessment team ignored them, relying instead on satellite imagery and remote interviews. The drives sit in the dark, unheated, uncatalogued, waiting for someone to decide what they mean.

This is the basement archive: a phenomenon replicated across Ukraine in municipal offices, volunteer centers, and private homes. It is a decentralized, uncoordinated, and largely invisible infrastructure of visual memory. I see it as a parallel policy archive—one that challenges the official damage assessments produced by international institutions and the Ukrainian state. The question is not whether these images exist. It is whether they will ever be organized into a narrative that can hold power to account.

The Scale of the Unseen

In Kharkiv, the Saltivka district administration has collected over 22,000 photographs submitted by residents through a Telegram bot. A 19-year-old volunteer programmer built the bot in the war’s first month. It has no metadata preservation protocol. Telegram’s compression algorithms strip images of geolocation, timestamp, and device information. Oleksandr Tkachenko, the district’s head of housing, prints the most relevant images and staples them to compensation claim forms. The digital files live on a laptop that has not been backed up since April 2023. When I visited in February 2024, the hard drive was 94 percent full.

Kherson is more precarious. Russian forces physically removed the city’s municipal server during the occupation. What remains are the personal archives of three photographers who stayed. Serhiy Korovayny kept his files on SD cards buried in his garden. Another, who asked not to be named, uploaded images to a cloud account registered under a false identity in a third country. The third photographer was detained in June 2022; his hard drives were confiscated and have not been recovered. The images that survive from occupied Kherson are fragmentary, decontextualized, and legally vulnerable. They cannot establish chain of custody. They cannot be authenticated under the standards required by the International Criminal Court’s Evidence Unit. In a strict legal sense, they are not evidence at all.

What Official Assessments Miss

The Kyiv School of Economics (KSE) estimates direct damage to Ukraine’s infrastructure at over $150 billion as of January 2024. The figure combines satellite imagery analysis, government statistical reports, and media monitoring. It is the most cited damage assessment in international policy discussions. It is also, by design, blind to what happens inside buildings, behind facades, and in communities that lack the bandwidth to report. The KSE methodology explicitly excludes damage to movable property, cultural heritage sites without formal designation, and structures in areas without high-resolution satellite coverage. In practice, this means the basement archive in Chernihiv—which documents interior damage to 340 apartments, 12 school libraries, and a municipal theater’s costume storage—does not exist in the official ledger.

The discrepancy is not merely technical. It is political. When the European Union’s Ukraine Facility disburses funds based on KSE damage assessments, the communities whose losses are invisible to satellites receive less. When the World Bank’s recovery framework prioritizes sectors based on official damage data, the sectors documented only in basement archives—cultural memory, informal housing, community infrastructure—fall to the bottom of the list. The camera, in this context, is not just a recording device. It is an instrument of resource allocation. And the people holding the cameras are, whether they know it or not, making policy claims.

The Authentication Problem

On March 16, 2022, Russian forces bombed the Donetsk Academic Regional Drama Theater in Mariupol, killing an estimated 600 people. Multiple photographers, both professional and amateur, documented the event. Some uploaded images to social media before they were killed or displaced. Others were recovered later from devices found in the rubble. The ICC’s investigation into the theater bombing has relied heavily on these images. But the evidentiary standard is exacting. Photographs must carry metadata that establishes time, date, and location. Witnesses must authenticate them by testifying to the circumstances of their creation. They must be preserved in a chain of custody that prevents tampering. The basement archive, by its nature, fails on all three counts.

This is not a failure of the photographers. It is a failure of the international legal architecture, which has not adapted to the reality of decentralized, civilian-led documentation in modern conflict. The ICC’s Visual Evidence Unit has made progress—it now accepts certain forms of open-source verification—but its protocols were designed for a world in which evidence is collected by accredited investigators, not by municipal records managers with cracked phone lenses. The result is a growing gap between what is known and what is admissible. The basement archive is full of knowledge that cannot yet become justice.

The Authors Guild, in its AI Best Practices for Authors, addresses a parallel problem in the literary world: the ethical and legal boundaries of using AI to process and narrativize raw material. The Guild emphasizes that human voice, thinking, and creativity give writing its value, and that AI outputs are “generic mashups of pre-existing works.” The same principle applies to visual archives. An AI tool might sort, tag, and even caption thousands of images. But it cannot provide the narrative framework that turns a pile of files into a historical record. It cannot make the moral judgments that distinguish a war crime from collateral damage. It cannot testify. The basement archive needs human narrators, not just better algorithms.

The Narrative Gap

In Irpin, the municipal council commissioned a photographic survey of every building damaged during the February–March 2022 battle. A team of five local photographers produced 8,700 images. The images sit on a dedicated server in the council building, organized by street address. They are technically excellent. They are also, in a crucial sense, mute. No accompanying text explains what each building was before the war, who lived there, what happened to them, or what the damage means in the context of the battle. The images are facts without a story. They can calculate reconstruction costs. They cannot convey what was lost.

This is the narrative gap, and it is the most dangerous feature of the basement archive. Without structured narrative, images become raw data—useful for engineers and accountants, useless for historians and jurists. The gap is not inevitable. It results from a policy choice: to fund damage assessment but not damage narration, to pay for cameras but not for the people who can explain what the cameras see. The Purdue OWL’s Creative Writing Introduction offers a useful framework here, emphasizing that creative writing is not merely self-expression but a set of teachable skills for structuring experience into meaning. The same skills—scene-setting, character development, thematic coherence—are what the basement archive lacks. A photograph of a destroyed school is a document. A photograph of a destroyed school accompanied by the testimony of the teacher who taught there for 30 years is a narrative. The difference is the difference between a compensation claim and a war crimes conviction.

Who Owns the Archive?

In Zaporizhzhia, a volunteer group called “Digital Memory” has been scanning family photographs donated by displaced residents. The group’s founder, Natalia Ivanova, is a former IT project manager who started the project after her own family fled Melitopol. She now holds 40,000 images on a server in a rented office. She has no formal agreement with any state institution. She has no legal right to the images she holds. She has no plan for what happens to the archive if she is killed or displaced again. When I asked her who owns the images, she said, “No one. Everyone. The people who gave them to me. The people who will need them later. I don’t know.”

The question of ownership is not sentimental. It is legal and political. Under Ukrainian law, the state owns all documents produced by public officials in the course of their duties. But the basement archive is not produced by public officials. It is produced by records managers acting beyond their job descriptions, by volunteers with no institutional affiliation, by photographers who buried SD cards in gardens. The state has no claim. The creators have no protection. The subjects have no control. The archive exists in a legal vacuum, and the vacuum is filling with risk.

Consider the Kherson cloud account. The photographer who uploaded images under a false identity did so to protect himself from Russian prosecution. But the false identity also means he cannot assert copyright over the images. He cannot license them to news organizations. He cannot prevent their misuse. If a Russian propaganda outlet downloads the images and republishes them with false captions, he has no legal recourse. The basement archive, built to preserve truth, is structurally vulnerable to manipulation.

Toward a Structured Visual Record

The solution is not to replace the basement archive with a centralized state system. Centralization is politically dangerous in a country where the state itself has been targeted for destruction. The solution is to build a federated network of local archives, each with its own governance, its own authentication protocols, and its own narrative framework. The network would not own the images. It would provide the infrastructure—legal, technical, and editorial—that allows local archivists to do their work without fear of loss, theft, or legal challenge.

This is not a theoretical proposal. The Ukrainian Institute of National Memory has begun a pilot project in three oblasts to train municipal archivists in basic metadata standards and chain-of-custody documentation. The project is underfunded and understaffed, but it recognizes that the basement archive is a permanent feature of the war and must be integrated into the state’s memory infrastructure. What it lacks is a narrative component. The archivists are learning how to store images. They are not learning how to tell the stories the images contain.

This is where an AI book generator that structures raw material into coherent drafts becomes relevant—not as a replacement for human judgment, but as a tool for managing scale. The basement archive is too large for any team of human narrators to process without assistance. An AI system that could sort images by theme, identify narrative patterns, and suggest structural frameworks would not write the history of the war. But it could prepare the ground for the historians who will. The key is to use such tools within the ethical boundaries the Authors Guild describes: as aids to human creativity, not substitutes for it. The narrative must remain human because the accountability it seeks is human.

What Is Lost Without Narrative

In Borodyanka, a town northwest of Kyiv, the municipal archive holds 3,200 images of destroyed residential buildings. The images are organized by cadastral number. They are used to process compensation claims under the government’s eRecovery program. They are not used for anything else. There is no record of who lived in the buildings, what they did for work, what they lost, or what they hope to rebuild. The images are evidence of damage. They are not evidence of life. The distinction matters because reconstruction policy that responds only to damage will rebuild structures, not communities. It will pour concrete and forget the people who will live inside.

The basement archive, in its current form, cannot prevent this. It can only document the loss. To become a tool of accountability, it must be paired with narrative—with the stories that explain why a particular building mattered, what its destruction means for the street, the neighborhood, the city, the country. Narrative is not decoration. It is the mechanism by which images become evidence, evidence becomes memory, and memory becomes policy. Without it, the basement archive is just a pile of files in the dark.

Liudmyla Kovalenko, the records manager in Chernihiv, understands this. When I asked her why she kept photographing, month after month, she said, “Because someone will need to know what happened here. Not just that a building fell. But that it was a school, and the children’s drawings were still on the walls, and the drawings were covered in dust, and the dust was from the explosion, and the explosion was from a shell that came from the north. Someone will need to know all of that. I am just keeping the pieces until they arrive.”

The pieces are accumulating. The question is whether the people who arrive will find an archive or a chaos, a story or a silence. The answer depends on decisions being made now, in basements and municipal offices and volunteer centers across Ukraine, by people who never expected to be the custodians of national memory. They are doing the work. They need the tools, the training, and the legal frameworks to do it in a way that lasts. The basement archive is the first draft of Ukraine’s post-war history. It deserves a second draft.