End-to-end encryption: what it does and does not protect
End-to-end encryption today is both a marketing phrase and a real technical guarantee. It is worth knowing which is which — because the difference between the two is whether the provider can read your messages.
“End-to-end encryption” (abbreviated E2EE) is one of the most frequently used and most frequently misunderstood terms. Let us look at what it actually covers — with no technical background needed.
The essence in a single sentence
The message is encoded by your device, and only the recipient's device can decrypt it. Anyone who gets access in between — the provider, the internet provider, an attacker on the network — sees nothing but unintelligible data.
The emphasis is on the “only the recipient's device” part. That is what distinguishes real E2EE from plain encryption.
What is the difference between “encrypted” and “end-to-end encrypted”?
Almost every service today encrypts at some level. The question is where the key is.
Transport encryption (the padlock in the browser's address bar): the message travels encrypted between your machine and the server. There, however, it is decrypted, processed, stored — and sent on to the recipient re-encrypted. So the provider sees the content. This is the default for most services.
End-to-end encryption: the key exists only on the two participants' devices. The server forwards the data packet, but has nothing to open it with. If a court asks it for the messages, it cannot hand over meaningful content, because it does not have it either.
A simple test to tell the two apart: if, when you log in to a service on a new device, your entire old message history is immediately there, then it is most likely not end-to-end encrypted. With real E2EE, recovering old messages takes a separate step, because the key was on the old device.
How does it work, very briefly?
Modern systems use a key pair. Everyone has a public and a private key:
- The public key can be distributed freely. Anyone can encode a message to you with it.
- The private key never leaves your device. Only it can decrypt what was encoded with the public key.
When you send a message, the app fetches the recipient's public key, encodes with it, and sends. The server merely forwards a meaningless heap of data.
Good systems go further than this: they generate a new key for every message and then discard it. This is called forward secrecy, and it ensures that even if someone later obtains your private key, they still cannot decrypt earlier conversations.
What E2EE protects against
- The provider. The operator cannot read your messages, cannot analyse them, and cannot hand over their contents.
- Data breaches. If the server is compromised, the attacker obtains an encrypted heap of data that is useless on its own.
- Network eavesdropping. Not even on public wifi can anyone see what you write.
- Official requests — in the sense that the provider cannot hand over what it does not have.
What it does not protect against
This is the more important part, because this is where most of the misconceptions are.
It does not constrain the recipient. Whatever you send, the other party does what they like with: copies it, forwards it, photographs it with another phone. E2EE protects the transfer; it is not a substitute for trust.
It does not protect your device. If someone gets access to your unlocked phone, they read everything — at that point the decryption has already happened. The same applies if spyware is running on the device: it sees the screen, and therefore the decrypted text.
It usually does not hide metadata. The content is secret, but often the fact of who communicates with whom, when and how often is not. That alone is frequently revealing. Anyone who wants to hide that as well needs further tools.
Screenshots cannot be reliably prevented. An app can flag that the other party took a screen capture, but no software will stop someone photographing the display with another device.
What to watch for when a service promises this
A few practical checkpoints:
- Is it on by default? If encryption has to be switched on separately, then most of your conversations will not be encrypted. In a good implementation there is no switch, because it is always active.
- Is there key verification? Serious systems let you compare the key fingerprint with the other party. Without this, it is hard to be sure you really are talking to who you think you are.
- What does the provider say about metadata? Precise, honest wording is a good sign. A “total anonymity” style promise with no specifics is suspicious.
- What happens when you switch devices? If everything comes across seamlessly everywhere, that is convenient — but then somewhere a key is being managed by the provider.
In summary
End-to-end encryption is real and important protection: it guarantees that nobody but you and the recipient can access the content of your message.
What it does not guarantee is that the recipient will be discreet, that your device is secure, or that the fact of the communication stays hidden. Technology protects the channel — the rest remains a matter of human decisions.