Every year, the average person takes over 2,000 photos. Most of them live in Google Photos or iCloud — two ecosystems that have become so frictionless, so deeply integrated into our phones, that opting out feels almost impossible.
But convenience has a cost. And in this case, it’s measured in something far more personal than money.
This piece explains exactly what Google and Apple do with your photos, why it matters more than most people realise, and why a genuinely independent alternative is not a luxury — it’s a necessity.
How Google Turns Your Memories Into Ad Revenue
Google Photos is “free.” Google is an advertising company. These two facts are not unrelated.
Image scanning and content understanding
Google’s terms of service grant the company a broad, royalty-free licence to analyse, reproduce, and create derivative works from content you upload — for the purpose of operating and improving their services. In practice, this means Google’s machine learning systems scan every photo you upload.
What does that scanning reveal? Far more than you might expect:
- Location data extracted from EXIF metadata tells Google everywhere you’ve been, with timestamps accurate to the second.
- Face recognition builds a model of who you spend time with — your family, your close friends, your habits.
- Scene classification identifies restaurants, gyms, beaches, hospitals, places of worship, and hundreds of other contexts.
- Object recognition reads product labels, car logos, clothing brands, and retail environments.
- Text recognition (OCR) extracts readable text from photos of documents, menus, boarding passes, and whiteboards.
This data flows into Google’s advertising infrastructure. When you search for holidays, Google already knows your travel history. When you browse for restaurants, Google knows where you ate last weekend. The connection is never explicit, but it is structural.
The “automated systems” clause
Google’s privacy policy states they use “automated systems” to “analyse your content.” This analysis is used to “personalise our services for you, including providing customised search results, personalised ads, and other features tailored to how you use our services.”
In plain language: the photo you took at a cancer screening clinic, or at your divorce lawyer’s office, or at a political rally — all of it flows into a system designed to make you more predictable and therefore more profitable to advertisers.
Account termination risk
Google has terminated thousands of accounts — often with no warning, no appeal, and no path to data recovery — for content flagged by automated systems. Parents have lost years of family photos after a legitimate medical image of a child was misclassified. Journalists have lost archives. Activists have lost evidence.
When everything you care about lives inside a single company’s ecosystem, their terms of service become your de facto law. And you have no vote on how those terms change.
What Apple Does That Most People Don’t Know
Apple markets itself as the privacy-first alternative to Google. Their hardware encryption and on-device processing are genuinely strong. But several practices complicate the clean narrative.
iCloud is not end-to-end encrypted by default
This is the most important thing most Apple users don’t know: photos stored in iCloud are not end-to-end encrypted unless you manually enable Advanced Data Protection — a setting that requires iOS 16.2 or later, a recovery contact or recovery key, and deliberate opt-in.
The default state? Apple holds the encryption keys. This means:
- Apple can comply with law enforcement requests for your photo library.
- Apple engineers with sufficient access can potentially view your content.
- A data breach at Apple’s infrastructure level could expose unencrypted copies of your photos.
Between 2019 and 2024, Apple received thousands of government requests for customer data across all markets. Unlike a service with true end-to-end encryption, they were technically capable of complying with many of them.
The CSAM scanning controversy
In 2021, Apple announced a system called CSAM Detection (Child Sexual Abuse Material Detection) that would scan photos on-device before uploading to iCloud. The proposal was withdrawn following intense backlash from cryptographers, privacy advocates, and civil liberties organisations worldwide.
The core objection was not about the stated goal — which everyone agreed was worthwhile — but about the precedent: a company building infrastructure to scan private content on your device, at scale, for a purpose defined by a list controlled by a third party. That same infrastructure, once built, could be repurposed. Apple paused the rollout but has never fully renounced the capability.
The episode revealed something important: even a company with genuinely strong privacy values has structural incentives and legal pressures that can override them. You cannot fully outsource the protection of your most private content to a third party.
The ecosystem lock-in mechanism
Apple’s photo export tools are deliberately cumbersome. Migrating a 10,000-photo library out of iCloud takes hours, requires specific software, and often loses metadata. This friction is not accidental. It’s a feature — from Apple’s perspective — that keeps you paying for iCloud storage every month.
When the cost of leaving is high enough, you stop asking whether you should leave.
Why “They Have Too Much Data Already” Is Not a Reason to Stop Caring
A common response to these concerns is fatalism: “They already know everything about me. What does it matter?”
It matters for several reasons.
The value of your photo archive compounds over time. A photo library from five years ago contains data about where you were, who you were with, what you owned, what you believed, what you looked like. As AI analysis improves, that historical data becomes increasingly valuable — both for targeting and, in adversarial contexts, for manipulation and coercion.
Context collapse is irreversible. Once your photos have been analysed by an advertising system, you cannot un-analyse them. The inference layer built from your content persists even after you delete the photos.
Legal protection lags badly. GDPR, CCPA, and similar frameworks were written for a world where data collection was slower and more visible. AI-powered inference from image data is largely unregulated, and the jurisdictions where Google and Apple store their data are not necessarily the ones with the strongest privacy laws.
The harm is often invisible until it isn’t. Most people never notice when their browsing behaviour changes because an advertiser bought an audience segment derived from their photo location data. The harm is diffuse, statistical, and deniable. That doesn’t make it less real.
What Independent Photo Storage Actually Means
“Independent” is not the same as “self-hosted.” Most people don’t want to run a server in their garage. What independence actually requires is:
- A provider whose business model does not depend on your data. If the product is storage, the revenue is storage. Full stop.
- Encryption architecture that does not require you to trust the provider. True end-to-end encryption means the provider cannot read your content even if compelled to.
- Transparent data handling. You should be able to see exactly what happens when you upload a file — no opaque “automated systems.”
- Portability on your terms. Exporting your data should be as easy as importing it.
- No training data extraction. Your photos should never be used to improve someone else’s AI.
How daftei Approaches This
daftei was built from the recognition that the companies offering “free” photo storage are not in the storage business. They are in the data business, and your memories are the inventory.
Encryption without compromise
Every file uploaded to daftei is encrypted in transit using TLS 1.3 — the same standard used by global banking infrastructure. At rest, all stored media and metadata is encrypted with AES-256. These are not marketing claims; they are the same standards applied to the most sensitive institutional data in the world.
We do not analyse your content for advertising
daftei does not build ad profiles. We do not scan your photos to infer your location history, your purchasing intent, your political beliefs, or your social graph. The AI features inside daftei — search, transcription, memory assistance — operate on your behalf, not on behalf of advertisers.
We will never sell your data. We will never train AI models for third parties using your content. We will never show you advertising based on your memories.
These are not aspirational statements. They are operational constraints built into how the company works.
Compliance as a floor, not a ceiling
daftei is GDPR and CCPA compliant. But we treat these frameworks as a baseline, not a goal. You can export your entire vault at any time from Settings. You can delete your account — and all associated data — instantly, with a 30-day grace window to change your mind. After Day 30, deletion is permanent and irreversible.
A business model aligned with your interests
daftei is a subscription product. When you pay for storage, that is the exchange in its entirety. There is no second transaction happening behind the scenes where your memories are the commodity. If daftei cannot justify its price with storage quality, reliability, and features, it should lose your business. That is the correct incentive structure.
The Practical Question
You might use Google Photos because it is excellent at face recognition, or because Shared Albums work across your family’s different phone types, or simply because changing habits is hard.
These are legitimate considerations. No product should ask you to sacrifice real utility for abstract principles.
But utility exists on a spectrum. The question is not “should I delete Google Photos today?” — it might be “where do the things that matter most belong?” Your most private moments, your children’s faces, the documents of your life — these deserve a home where the host has no financial incentive to analyse them.
The second question is: “what happens if my Google account gets terminated tomorrow?” For most people, the answer is: I lose everything.
That is a structural fragility worth taking seriously, regardless of your position on data privacy in the abstract.
A Different Kind of Vault
The word “vault” is intentional. A vault is a place where what matters most is protected — not from you, but on your behalf. Where access is controlled by the owner, not the institution.
daftei is that vault. It stores your photos, voice notes, documents, and journeys. It uses AI to help you search and rediscover them. It does not sell what it learns to someone else.
Your memories are yours. The infrastructure that holds them should act accordingly.