In July 2026, the Defense Manpower Data Center — the US military agency responsible for maintaining personnel records for service members and their dependents — discovered it had a problem. An unauthorized actor had accessed a database containing records for approximately three million military personnel.
The breach itself is serious. What made it worse: the data was stored completely unencrypted.
Social Security numbers, dates of birth, contact information — all of it sitting in a database with no encryption layer. The data was accessed for roughly nine months, from October 2025 until the discovery in July 2026. During that window, the intruder had unrestricted access to the raw data.
What “Completely Unencrypted” Actually Means
There’s a meaningful difference between a breach of encrypted data and a breach of unencrypted data, and it matters for how you think about your own digital records.
When data is encrypted at rest — stored in encrypted form on disk — an attacker who gains access to the database doesn’t automatically gain access to the data. They get an unreadable sequence of bytes that requires a key to decrypt. Without the key, the data is computationally infeasible to use.
When data is stored unencrypted, an attacker who can read the database reads the data. There’s no second line of defense. The access control layer — usernames, passwords, access controls — is the only barrier. If that layer fails, the data is fully exposed.
In the DMDC breach, the access controls failed, and there was no encryption layer to fall back on. For nine months, approximately three million people’s sensitive personal records were readable in plain form.
Why This Happens in Large Organizations
The DMDC breach is not an isolated failure mode. Storing sensitive data unencrypted is more common than most organizations would like to admit, and the reasons are instructive.
Legacy systems. Many large organizations — governments included — run databases that were built before encryption at rest was standard practice. Retrofitting encryption into a production system is technically complex and operationally risky. Teams defer the work, document it as a known gap, and accept the risk. Years pass.
Operational complexity. Encryption at rest adds management overhead: key management, rotation schedules, access to decryption infrastructure. For teams focused on availability and functionality, encryption is often deprioritized in favor of keeping the system running.
Assumption of perimeter security. Organizations often reason that if their access controls are strong, the data never needs encryption because unauthorized users can’t reach it. This assumption fails when an authorized user account is compromised, when insider threats emerge, or when perimeter controls have a gap that wasn’t anticipated.
Compliance without substance. Some organizations treat encryption as a checkbox — implementing it at the application layer for the data they know auditors will examine, while leaving other data stores unencrypted. This creates a false sense of assurance while leaving significant gaps.
The result is databases that are protected by a single layer of authentication, with no backup protection if that layer is breached.
The Asymmetry of Sensitive Personal Data
The DMDC breach involved Social Security numbers, dates of birth, and contact information. For the three million people affected, those aren’t just administrative records — they’re the raw material for identity theft.
Social Security numbers are particularly problematic because they’re static. You can change your email address, your phone number, and your password. You cannot change your Social Security number. When it’s exposed, it remains exposed. The window of risk for the affected individuals doesn’t close with the breach discovery — it persists for years.
This is the same asymmetry that makes biometric data — faces, fingerprints, iris scans — especially sensitive. You can’t issue replacement faces. The personal identifiers that matter most are precisely the ones that can’t be rotated when they’re compromised.
For military personnel specifically, a compromise of this kind carries additional risks: the data could be used to target or blackmail individuals with security clearances, to identify and approach relatives of service members, or to generate fraudulent credentials and identities that are harder to detect because they’re built on real military records.
What AES-256 Encryption at Rest Actually Provides
The relevant standard for encrypting data at rest in US federal systems is AES-256 — Advanced Encryption Standard with a 256-bit key — operating in an approved mode. This is the same standard required under FIPS 140-2 and its successor FIPS 140-3 for federal cryptographic systems.
AES-256 encryption at rest means that data stored on disk is encrypted. When an application needs to read the data, the system decrypts it using a key stored separately from the data itself. An attacker who gains access to the storage medium or the database file — without access to the key — cannot read the data.
The protection isn’t absolute. An attacker who also compromises the key management system, or who gains access through an authenticated application-layer connection, can still reach the decrypted data. Encryption at rest is a defense in depth measure, not a complete solution.
But it’s a significant one. In the DMDC case, if the data had been encrypted at rest with AES-256, the nine-month unauthorized access would likely have yielded nothing useful — a large file of cipher text rather than 3 million exposed records.
Personal Files Deserve the Same Standard
The DMDC breach involved a government agency with legal and regulatory obligations to protect sensitive data. But the lesson applies just as directly to how individuals store their own personal records.
Most people have a digital archive of sensitive documents: tax returns, medical records, insurance documents, legal paperwork, passport and ID photos, bank statements. These are stored in various places — local drives, email attachments, shared cloud folders, photos apps.
Many of these storage locations have no encryption at rest. Or they encrypt at rest in ways that the platform controls entirely — meaning the platform can decrypt your data without your knowledge, for a variety of purposes including responding to legal requests, data analysis, or improving AI features.
The standard to look for when choosing where to store sensitive personal files is encryption at rest with strong algorithms — specifically AES-256. This doesn’t require the data to be end-to-end encrypted (which adds key management complexity for the user), but it does require that data stored on disk is not readable in plain form.
daftei applies AES-256 encryption to all files stored at rest. Stored data is not analyzed for AI training, not shared with advertisers, and accessible from iOS, Android, and web. The distinction between server-side AES-256 and end-to-end encryption is worth understanding clearly: daftei provides the former, which means the service can decrypt data to provide it to you on request, but your data is not stored in plain form on disk, and is not accessible to anyone who gains unauthorized access to the underlying storage without also compromising the decryption infrastructure.
What You Can Do With Your Own Records
The DMDC breach is a reminder that you don’t fully control how organizations that hold your data protect it. You can, however, control the records you keep yourself.
Audit what you’re storing and where. Go through your cloud storage folders, your email attachments, and your phone’s files app. Identify where sensitive documents are living — particularly Social Security information, passport copies, medical records, and financial documents.
Check your storage provider’s encryption practices. Look for explicit language in the privacy policy or security documentation about encryption at rest. “Encrypted in transit” (TLS) is not the same as “encrypted at rest.” You want both, but it’s the at-rest encryption that matters if someone accesses the stored files.
Don’t store sensitive documents in unprotected locations. Email attachments, shared Google Drive folders with public link sharing, and screenshots in a camera roll that syncs to a cloud service are common but insecure homes for sensitive documents.
Use a purpose-built storage service for your most sensitive files. General-purpose cloud storage platforms — Google Drive, Dropbox, iCloud — are designed for collaboration and accessibility. Purpose-built private storage services are designed for confidentiality. The design goals produce different defaults.
Think about what happens when data is breached. Encrypted data, if breached, is largely useless to an attacker. Unencrypted data, if breached, is immediately readable. When choosing where to store sensitive personal records, ask not just “is this convenient?” but “what happens if this system is compromised while my data is in it?”
The Standard Should Be Higher
The DMDC breach is, in a narrow sense, a government IT failure. But the broader failure it illustrates — treating encryption at rest as optional rather than mandatory for sensitive personal data — is widespread across both public and private sectors.
Federal standards like FISMA require agencies to implement encryption for sensitive data. That requirement apparently didn’t prevent the DMDC from leaving three million records unencrypted for years. Standards without enforcement and accountability don’t protect data.
For individuals, the lesson is simpler: encryption at rest is not a premium feature. It’s a baseline expectation for any service that stores your sensitive personal records. If a service doesn’t apply it, your files are only as safe as the service’s access controls — and access controls fail.
Sources: Privacy Guides – Data Breach Roundup Sep 25–Oct 1, 2026