End-to-End Encryption, Explained Properly

The sealed envelope: what end-to-end actually means

Imagine you want to send a letter to a friend. You write it, seal it inside an envelope, and drop it at the post office. The postal workers carry it across the country. They can see the address on the front. They can see when it was posted and when it arrived. But they cannot read what is inside, because the envelope is sealed and only your friend has the means to open it properly.

End-to-end encryption works on exactly that principle. Your message is scrambled, mathematically, on your device before it leaves. It travels across the internet in a form that looks like nonsense to anyone who intercepts it. When it arrives at the other end, only the recipient's device holds the key to unscramble it. Not the company whose app you used. Not your internet provider. Not a government agency tapping the line. Just the two of you.

The phrase "end-to-end" is the important part. It means the encryption holds from one end of the conversation to the other. There is no point in the middle where the message sits in readable form on someone else's server. Some older systems encrypted the connection between you and the company's server, then decrypted it there, then re-encrypted it to send on. That is not end-to-end. The company in the middle could read everything. End-to-end encryption closes that gap.

The technical machinery behind this involves something called public-key cryptography. Each person in a conversation has two mathematically linked keys: a public one, which anyone can use to lock a message for you, and a private one, which only your device holds and which is the only thing that can unlock it. Think of the public key as a padlock you hand out freely. Anyone can click it shut around a message. But only you have the key that opens it. The maths involved means that working backwards from the locked message to guess the private key would take longer than the age of the universe on any computer that currently exists.

Practical takeaway: You do not need to understand the maths. What matters is the principle: if an app offers genuine end-to-end encryption, the company running it genuinely cannot read your messages, even if a court orders them to hand over the contents.

Encryption you already use every day without noticing

Most people assume encryption is something for journalists, activists, or people with something to hide. It is not. You almost certainly use it several times a day without a second thought.

When you log in to your bank, the padlock icon in your browser's address bar and the "https" at the start of the URL tell you that your connection is encrypted. That protects your password and account details from anyone watching the traffic between you and the bank. This is not quite end-to-end encryption in the messaging sense, because the bank's server does decrypt your data to process it, but it is strong encryption protecting you in transit.

Messaging is where true end-to-end encryption has become genuinely widespread. WhatsApp has applied it to all conversations by default since 2016. Apple's iMessage uses it when both parties are on Apple devices (the blue bubbles, not the green ones). Signal is built around it entirely. FaceTime calls are end-to-end encrypted. Even some features inside Facebook Messenger now offer it, though the rollout and defaults there have been more complicated, and it is worth being aware of what messaging apps inside social platforms collect beyond the message content itself.

These are not recommendations. They are simply the places where most UK adults already encounter this technology, often without realising it. The point is that end-to-end encryption is not exotic. It is already woven into ordinary daily life.

Email is a notable exception. Standard email, the kind sent between different providers, is generally not end-to-end encrypted. The providers can read it. There are tools that add encryption to email, but they require both parties to set them up, which is why they have never become mainstream. For most people, email remains a postcard rather than a sealed envelope.

Practical takeaway: Check whether the messaging apps you use most often have end-to-end encryption switched on by default. Most major apps will say so clearly in their settings or help pages. If you are not sure, searching the app's name alongside "end-to-end encryption" on its official support site will give you a straight answer.

What E2EE protects you from, in real situations

Abstract security is hard to care about. So here is what end-to-end encryption actually prevents, in situations that are genuinely relevant to ordinary people in the UK.

First, it protects you from data breaches at the company level. Companies get hacked. It happens regularly and to organisations that ought to know better. If a messaging platform stores your messages in readable form on its servers and those servers are breached, your messages are exposed. If the messages were end-to-end encrypted and the company never held the keys, there is nothing useful to steal. The attacker gets scrambled data.

Second, it protects you from overly curious insiders. Employees at large technology companies do not generally spend their days reading your messages, but the access has existed in systems without proper end-to-end encryption. Researchers and journalists have documented cases where employee access to private communications was broader than users expected. End-to-end encryption removes that possibility structurally, not just by policy.

Third, it protects the content of your communications from being handed to third parties in response to legal requests. If a company receives a court order demanding the contents of your messages and those messages are genuinely end-to-end encrypted, the company cannot comply, because it does not have the contents. It can hand over what it does hold (more on that shortly), but not the words you wrote.

This matters for people in all sorts of ordinary situations. A person discussing a medical condition with a family member. Someone seeking advice about a difficult workplace situation. A journalist's source. A domestic abuse survivor communicating with a support worker. None of these people are criminals. All of them have legitimate reasons to want their words to stay private.

Fourth, on public or shared Wi-Fi networks, such as in a café, a hotel, or an airport, your internet traffic passes through infrastructure you did not choose and do not control. End-to-end encrypted messages remain protected even there, because the scrambling happened on your device before the data hit the network.

Practical takeaway: If you have a conversation coming up that you would prefer to keep genuinely private, checking whether the app you plan to use offers end-to-end encryption by default is a reasonable and proportionate thing to do. It takes about thirty seconds.

What it does not hide: the envelope has a name and a postmark

Here is where a lot of explanations of encryption go quiet, and where the sealed-envelope analogy becomes especially useful.

Go back to that envelope. The postal worker cannot read what is inside. But they can see the name and address on the front. They can see when it was posted. They can see how heavy it is, roughly, and whether it came from London or Edinburgh. All of that information exists outside the sealed part.

In the world of encrypted messaging, this outer information is called metadata. And end-to-end encryption does not protect it.

The company running your messaging app typically knows: who you messaged, when you messaged them, how often you message that person, roughly how long your messages are, your IP address (which can indicate your location), and what device you used. All of this is available to the company even when it cannot read a single word you wrote.

Metadata is more revealing than most people realise. A former director of the US National Security Agency once said, publicly, that the agency could kill people based on metadata alone. That is an extreme illustration, but the underlying point is serious. Knowing that you called a cancer helpline three times last week, then called a solicitor, then messaged a family member for two hours, tells a story even without the words.

In the UK, under the Investigatory Powers Act 2016, authorities can require communications providers to retain and disclose certain categories of metadata. The legislation is publicly available on GOV.UK if you want to read the framework. The content of an end-to-end encrypted message may be beyond reach. The fact that you sent it, and to whom, may not be.

This is not a reason to distrust encryption. It remains enormously valuable. But it is a reason to understand what it does and does not promise. The content of your conversation is sealed. The envelope still has your name on it. If you want to understand how this kind of metadata feeds into broader tracking, the picture of how companies track you online is worth reading alongside this piece.

Practical takeaway: If the fact of a conversation is itself sensitive, not just its content, end-to-end encryption alone is not a complete answer. Understanding what metadata your app collects is a separate question from whether your messages are encrypted.

The backup trap: how a cloud copy can undo the promise

This is the part that catches people out most often. It is quiet, it is easy to miss, and it matters a great deal.

Many messaging apps offer to back up your conversation history to a cloud service. On an iPhone, that typically means iCloud. On Android, it has often meant Google Drive. The backup exists so that if you lose your phone or switch to a new one, you can restore your messages. That is genuinely useful.

But here is the problem. When your end-to-end encrypted messages are copied into a standard cloud backup, they may be stored in a form that the cloud provider can access. The encryption that protected your messages in transit does not automatically extend to the backup copy sitting in someone else's storage. The sealed envelope was opened on your device, the letter was read, and then a photocopy was made and left in an unlocked filing cabinet.

WhatsApp, for a long time, backed up to Google Drive or iCloud without encrypting the backup end-to-end. The messages were protected in transit but potentially readable in the backup. Meta introduced end-to-end encrypted backups as an option in 2021, but it is not always the default, and it requires the user to set it up deliberately. Apple's iMessage backs up to iCloud, and if iCloud Backup is turned on, Apple holds a key that could be used to decrypt those messages. Apple's own security documentation explains the distinction between iCloud Backup and iCloud's end-to-end encrypted data categories, and it is more nuanced than most people expect.

The practical consequence is real. Researchers and legal cases have shown that law enforcement requests for cloud backups have produced message contents that the messaging app itself could not have provided, precisely because the backup was not end-to-end encrypted. The app kept the promise. The backup broke it.

This is not a reason to disable backups entirely. Losing years of messages because your phone was stolen is a genuine cost. But it is a reason to understand what your backup settings actually do, and to make a conscious choice rather than accepting whatever the default happens to be.

Practical takeaway: Go into your messaging app's settings and find the backup or chat history section. Check whether end-to-end encrypted backup is available and whether it is turned on. If it is not an option, you should know that your backup may be accessible to the cloud provider, regardless of how the app itself encrypts your messages.

Why governments keep arguing about it, in one calm section

You may have seen headlines about governments wanting to weaken or ban end-to-end encryption. The argument comes up regularly in the UK, in the EU, and in the US. It is worth understanding what is actually being debated, without getting pulled into the politics of it.

The core tension is genuine. End-to-end encryption protects everyone equally. It does not distinguish between a domestic abuse survivor and someone planning serious harm. Law enforcement agencies argue, with some legitimacy, that encryption can make it genuinely impossible to investigate certain crimes, particularly those involving child sexual abuse material. That is a serious concern, not a manufactured one.

The technical counter-argument, made consistently by cryptographers and security researchers including those at the UK's National Cyber Security Centre and the Electronic Frontier Foundation, is that there is no known way to build a backdoor that only the right people can use. A mechanism that allows a government to access encrypted communications would, by its nature, also be a vulnerability that hostile states, criminals, and other actors could potentially exploit. You cannot build a door that only opens for the police. The lock does not know who is knocking.

The UK's Online Safety Act 2023 included provisions that raised significant concern among security experts, because of clauses that could in principle require platforms to scan message contents, which is incompatible with genuine end-to-end encryption. Several major platforms indicated they would rather leave the UK market than compromise their encryption. The Information Commissioner's Office has separately emphasised that strong encryption is a tool that protects UK citizens' privacy rights under UK GDPR.

As of the time of writing, the most contentious provisions of the Online Safety Act have not been activated, and the debate continues. This is not a settled area.

The point for an ordinary reader is simply this: when you hear that a government wants to weaken encryption, the question to ask is not "what do they want to read?" but "what would the world look like if every sealed envelope could be opened by the right authority?" The answer to that question is not straightforward, and reasonable people disagree about the trade-offs. But the technical reality, that a backdoor for the good guys is a backdoor for everyone, is not really disputed by anyone who has studied the engineering.

Practical takeaway: Following this debate through primary sources, rather than through newspaper headlines, gives a much clearer picture. The NCSC's published guidance on encryption and the ICO's statements on privacy-enhancing technologies are both publicly available and written in accessible language.

Checking your own apps: what to look for this week

None of this is useful unless it connects to something you can actually do. Here is a short, practical checklist. It should take you no more than fifteen minutes.

1. Find out which of your messaging apps use E2EE by default

Open each app you use regularly and look in Settings, then Privacy or Security. Most apps that offer end-to-end encryption will say so explicitly. If you cannot find it in settings, the app's official help pages will tell you. What you are looking for is whether E2EE is on by default for all conversations, optional, or not available at all.

2. Check whether secret or private modes require manual activation

Some apps offer end-to-end encryption only in a specific mode. Facebook Messenger's end-to-end encrypted conversations, for example, have historically been a separate thread type rather than the default for all messages. If encryption requires you to start a special kind of conversation, your regular chats may not be protected in the same way.

3. Look at your backup settings

In WhatsApp, go to Settings, then Chats, then Chat Backup, and look for the end-to-end encrypted backup option. In iMessage, consider what iCloud Backup means for your message history, and read Apple's support documentation if you want to understand the detail. The key question is: if the cloud provider received a legal request for your backup, could they hand over readable message contents?

4. Think about email separately

Standard email between different providers is almost certainly not end-to-end encrypted. If you regularly send sensitive information by email, it is worth knowing that. Encrypted email exists but requires both parties to set it up. For most people, the practical answer is to use a messaging app with E2EE for anything genuinely sensitive, and treat email more like a postcard.

5. Remember what encryption cannot do

End-to-end encryption protects the content of your messages. It does not hide the fact that you sent them, who you sent them to, or when. If you need that level of privacy, that is a different and more complex question, and it goes beyond what any single app setting can solve.

The goal here is not to make you anxious about every message you send. Most conversations do not need extraordinary protection. But understanding what your tools actually do, rather than assuming they do more or less than they do, puts you in a genuinely better position. That is all this is about.

End-to-end encryption is one of the most important privacy technologies in everyday use. It works. It has real limits. And now you understand both.