314 MEPs voted to stop Chat Control. It passed anyway. That is a big problem

314 MEPs voted to stop Chat Control. It passed anyway. That is a big problem

“314 MEPs voted to stop Chat Control. 276 voted to keep it. 361 votes were needed to reject it. The majority lost.”

On 9 July 2026, the European Parliament voted on whether to block the return of Chat Control 1.0.

Three hundred and fourteen MEPs voted to reject it, while two hundred and seventy-six voted in favor of keeping it. The threshold required to stop the law was 361, meaning that the law passed.

There’s no mistake or obscure parliamentary procedure at play here. The majority of MEPs who voted opposed the measure, but it passed anyway.

In this guide, you’ll learn more about how this happened and what it means in practice for Internet users in the EU.

What Chat Control 1.0 actually was

Regulation (EU) 2021/1232 was what’s known as a “temporary derogation” from the ePrivacy Directive.

In other words, it was not a “Snooper’s Charter” in the same way as the UK’s Investigatory Powers Act and did not impose an EU-wide requirement to scan everyone’s private communications.

Its real purpose was to provide legal permission for providers of electronic communication services, who may voluntarily scan for known child sexual abuse material (CSAM) without violating EU privacy rules that would otherwise forbid these kinds of practices.

The operative word here is “may”. While Regulation 2021/1232 permitted scanning for CSAM and signs of child solicitation, subject to safeguards, it did not require them to do so.

It also applied mainly to services that process unencrypted messages, or those that were encrypted only on the server-side. Services that actually used it included Gmail, Facebook Messenger (before the platform enabled end-to-end encryption), Instagram DMs, Skype, Snapchat, iCloud Mail, and Xbox.

Services that did not (and technically could not) scan messages under this framework included WhatsApp, Signal, and iMessage with iCloud backups disabled.

Regulation 2021/1232 did not require providers to weaken or circumvent end-to-end encryption. It expired on 3 April 2026 when negotiations between the European Parliament and the Council failed to produce an agreement on a further extension.

Comparison table of Chat Control 1.0 vs 2.0 showing status, scanning type, encryption impact, and scope of detection

The part that almost everyone gets wrong

Although most recent media coverage has conflated the two, Chat Control 1.0 and Chat Control 2.0 are not the same law, though this confusion has led some readers in the EU to believe that their WhatsApp messages are already being read by the authorities.

The two proposed laws are distinct legislative instruments with different legal bases, different mechanisms, and radically different implications.

1.0 is the law that was passed. 2.0 is the law that matters

Chat Control 1.0 is the temporary ePrivacy derogation whose proposed extension was recently debated again. It covered platforms that already have access to message content, but didn’t touch end-to-end encrypted communications directly.

Chat Control 2.0, formerly the Child Sexual Abuse Regulation (CSAR), is the proposed permanent framework that would introduce mandatory detection orders. This would potentially require communications providers, including those offering E2EE, to scan content either server-side or, in the most contested versions, through client-side scanning on your device before the message is encrypted. This is the bill that has faced strong opposition from privacy advocates like the EFF.

The legislation remains under negotiation and hasn’t passed at the time of writing (July 2026).

The EU Parliament’s position currently excludes end-to-end encrypted communications from detection orders, while the EU Council’s latest negotiating position has moved away from mandatory detection orders in favor of allowing voluntary scanning on a permanent basis.

The confusion between Chat Control 1.0 and Chat Control 2.0 still matters, though, as version 1.0 never required platforms like WhatsApp to weaken end-to-end encryption or scan messages before encryption.

The greater potential danger to EU Internet privacy will lie in the eventual wording and implementation of Chat Control 2.0, should it become law.

Five rounds and a procedural move

The surprising July vote that we mentioned at the start of this guide was the result of a sequence that has been described by certain privacy advocates as a political ‘maneuver’ rather than a legitimately debated policy.

In March 2026, the European Parliament voted to support extending Chat Control 1.0 until August 2027, but only with additional safeguards.

The EU Parliament’s position would have excluded end-to-end encrypted communications, limited scanning largely to known or previously flagged CSAM, and required measures to target users or defined groups identified by a judicial authority who were reasonably suspected of having links to such material.

However, the EU Parliament and  Council failed to reach an agreement on the extension during negotiations. As a result, Regulation 2021/1232 expired on 3 April 2026.

Then something rather unorthodox happened: EU Parliament President Roberta Metsola raised the file at the European Council summit in June, urging leaders to push the interim regulation forward despite the Parliament having already rejected it.

The EU Council then adopted a new first-reading position, effectively resurrecting the same text Parliament had buried but using a procedure where members’ silence can be interpreted as approval.

In other words, the second reading would have required an absolute majority of the entire Parliament to reject the Council’s position (361 of 720 MEPs), not just the majority of those who show up to vote on the day.

As the vote was scheduled during the last sitting before the summer recess, when the chamber is at its emptiest, it’s easy to see why some privacy advocates might feel that the vote was engineered to an extent.

The fifth trilogue round on Chat Control 2.0 (the permanent framework) took place on 29 June 2026. It collapsed without agreement, particularly over the question of suspicionless scanning. A sixth round is expected in September 2026 under the Irish EU presidency, which has listed the file as a priority and has historically been supportive of client-side scanning of the kind outlined in Chat Control 2.0.

However, the most uncomfortable point that privacy advocates have been making is that reinstating Chat Control 1.0 removed the Council’s structural incentive to compromise on Chat Control 2.0.

To put it plainly, if scanning is allowed to continue, even under a technically voluntary and legally contested framework, there’s little political urgency to reach a deal on the permanent regulation behind Chat Control 2.0.

Does Chat Control work?

The short answer is: probably not. The European Commission’s own implementation report, published in November 2025 (as required by the regulation), couldn’t establish whether the ‘temporary derogation’ had a measurable effect on the detection or prosecution of abuse.

The report also concluded that the currently available data was insufficient to determine whether the measure achieved its stated balance between combating crime while protecting EU Internet users’ fundamental rights.

There are also issues surrounding accuracy. The Commission’s report  found that AI classifiers used to flag unknown material produced false positives in up to 20% of cases. Classification based on digital hashes of known CSAM was, naturally, more accurate.

Germany’s Federal Criminal Police Office, the BKA, also found that approximately 48 percent of Chat Control alerts forwarded from NCMEC in 2024 were not criminally relevant. In other words, the alerts didn’t directly point to illegal material. Roughly 99,375 non-actionable reports out of 205,728 were received, and this rate has been rising year-on-year.

According to former MEP Patrick Breyer, an estimated 99% of the material reported by Meta under Chat Control was already known to the authorities, making it virtually useless for stopping ongoing and new abuse.

According to Breyer, the EU Commission has also been unable to provide any evidence that indiscriminate scanning of private communications has resulted in more convictions.

He has also stated that 40% of investigations triggered by Chat Control flags actually target minors themselves, not perpetrators. For instance, children sharing images of themselves can become the subjects of criminal investigations because an algorithm has flagged the content.

The EU Commission and advocates for the European People’s Party (EPP) have argued that voluntary detection is currently the only mechanism producing CSAM reports at a meaningful scale in Europe.

They also claim that individual companies can change their data sharing and privacy policies overnight, and that there is no equivalent EU infrastructure to replace US-based reporting. From this perspective, permitting derogation to lapse weakens the ability to identify ongoing abuse and support victims. Certain child protection organizations have backed the derogation on these grounds, not just law enforcement agencies.

The counterargument

Nobody in this debate is arguing that CSAM should go undetected or unpunished. The counterargument is whether suspicionless scanning of all Internet users is a necessary and proportionate means to detect it.

The question also remains as to whether targeted scanning under judicial authorisation (applied to specific individuals on the basis of reasonable suspicion) would necessarily produce better intelligence at a lower cost to the privacy of the vast majority of users who are not criminals.

The EU Council’s own Legal Service reportedly warned in June 2026 that generalized “voluntary” scanning may be incompatible with Article 7 of the EU Charter: the right to private and family life, absent any reasonable suspicion by law enforcement or prior judicial authorization. These aren’t the opinions of fringe crypto-anarchists, but the Council’s own lawyers.

What this means for EU Internet users in concrete terms

Unencrypted/server-side encrypted content

If you use Gmail, Instagram DMs, Snapchat, Cloud Mail, or Xbox messaging, then technically your unencrypted content can be scanned against hash databases of known CSAM, without a warrant and without any individual suspicion attaching to you.

E2EE apps like Signal or WhatsApp

Message content for secure messaging apps that use end-to-end encryption like Signal and WhatsApp isn’t covered under the current law. The E2EE exemption amendment passed by the EU Parliament, if the EU Council accepts it, states this specifically. Metadata and certain backup data are a separate issue (see below).

Message history backups

In some cases, the storage methods used by platforms like Google and iCloud Drive can mean that backups of your message history sit outside the encryption envelope.

These cloud copies are not automatically protected by WhatsApp’s normal message-level end-to-end encryption. Unless the user enables WhatsApp’s separate end-to-end encrypted backup feature, the backup may instead rely on the cloud provider’s standard encryption and account security.

For instance, if you choose to back up your WhatsApp message history to iCloud unencrypted and don’t already have Apple’s “Advanced Data Protection” enabled, then theoretically your WhatsApp messages could be exposed via these backups, even if the original method used to encrypt them used end-to-end encryption.

The risk of false positives

As we’ve learned, a significant number of Chat Control alerts don’t reveal any illegal material and may even target children themselves in some cases. Nevertheless, a false positive will still likely result in a generated report in this context.

It could lead to questioning by law enforcement, as well as a search warrant and possible device seizure. Even if nothing illegal is found, innocent people could suffer reputational damage if people within their community discover that their device was searched for CSAM in the first place.

Timeline showing how Chat Control returned, from the 2021 derogation to Parliament's failed vote on 9 July 2026

Your practical stack: in order of impact

Use a secure E2EE messenger.

In this situation, the safest end-to-end encrypted messengers are open-source projects like Signal, Matrix, and Session. As the client source code is publicly available for such apps, it can regularly be reviewed by security experts to check for security bugs or coded backdoors.

Most messaging programs of this kind offer you a way to view the ‘public’ encryption keys used for messages and calls to reduce the possibility of a “Man In the Middle” (MITM) attack.

For instance, on Signal, tap a contact name and choose ‘View Safety Number’. Compare the 60-digit number or scan the QR code on your screen using your friend’s device, ideally in person.

Configure or disable cloud backups of your message history

This is a vital security gap that using end-to-end encryption apps doesn’t close in itself. As we outlined above, if you choose to save unencrypted backups of your message history to a third-party storage service like iCloud or Google Drive, then it could potentially be exposed to unauthorized parties.

The best solution is to choose a secure messaging app that supports end-to-end encryption of backups natively. For example, in Signal, if you go to Settings > Backup, you can start backing up your messages and media using end-to-end encryption via the platform itself.

If you choose this, the app will generate a 64-character recovery key. Write this down and put it in a safe place, like a secure password manager, as you’ll need it to recover messages on a new device. The recovery key itself isn’t stored in plaintext on Signal’s servers, so the platform itself can’t view your message history.

Encrypt what you store, not just what you send.

As we’ve seen already, third-party services that either don’t encrypt files you upload or only encrypt them server-side could technically be compelled to provide your sensitive information under regulations of the kind outlined by Chat Control 2.0.

The best defense against this is to encrypt files locally before uploading them to the cloud. Reputable open-source utilities like Cryptomator, Gocryptfs, and Veracrypt can allocate an area to place unencrypted files on your device before encrypting and syncing them to cloud storage platforms like Google Drive.

When choosing a passphrase, use a combination of at least six random words, e.g. by using Diceware, and note it down in a safe place.

Understand which jurisdiction holds the data you generate.

EU-based providers operate under different rules than those based in, say, the USA, so the infrastructure routing of any of your reports will differ accordingly. While there’s no such thing as a privacy-perfect jurisdiction that will never reveal any sensitive data under any circumstances, pay close attention both to where a provider is headquartered and where its physical infrastructure is based.

Network layer privacy

While we will discuss this in more detail in the next section, remember that many web apps could technically be targeted under Chat Control 2.0, including web browsers. For best protection, use an open-source, privacy-focused browser like LibreWolf or Brave. You’ll learn more about what network layer privacy steps you can take in the next section.

What a VPN can do against Chat Control - and what it can’t

In the worst-case scenario we’ve been discussing throughout this guide, a VPN can’t in itself stop Chat Control.

If a message were to be scanned, this would happen on the platform’s servers, or in the case of client-side scanning, on the device before information is transmitted.

While a VPN can move data over an encrypted virtual tunnel between your device and the server, this protection doesn’t extend to either of these surfaces. These kinds of scans happen at the application layer, while a VPN operates at the transport layer. That means if a VPN provider is claiming that it offers reliable protection against Chat Control, you should be very suspicious of its claims.

In this context, what a VPN can do includes:

Separating your network identity from your traffic

As your sensitive web traffic is encrypted, there’s no automatic way for your ISP or local network operator to see which sites you’re visiting or correlate your activity across them. The VPN connection itself can usually be detected, as well as the destination IP of the VPN server.

Removing your IP address as the correlation key

As your traffic is routed through the VPN server, your public IP address will appear to match that of the server, not your device. This makes it much harder for platforms to track and correlate your activity across different services and time periods using IP-based tracking methods. Naturally, masking your IP address using a VPN has no significant effect on non-IP-based types of tracking, such as browser fingerprinting.

Places your VPN traffic in a jurisdiction you choose

If you choose to connect to a VPN server in a different country, then you can usually access content on the same terms as someone in that location, and be subject to the same laws and regulations as someone who’s physically present there. While this doesn’t change the reality of where you and your router are based, it can make access by authorities in another jurisdiction less direct, as they may have to make cross-border legal requests and show justification for needing such information.

Matters more, not less, as the metadata layer becomes the surveillance layer

As the legislative pressure on content encryption increases within the EU, the metadata layer, such as who you’ve spoken with and when, will likely become an increasingly valuable surveillance surface. In such an environment, this means network-layer privacy  of the kind offered by a VPN matters more.

Still, the fairest conclusion to draw from recent developments is that while content encryption is being legislated against in the EU to some degree, transport encryption is not. Both are part of the same defensive posture in protecting your Internet privacy, and neither is a substitute for the other.

The correct response to privacy threats like Chat Control is not to choose one option but to apply both at the layers where they’re most effective.

The vote was procedural, the fight is not.

The significance of the July 2026 vote is less about what Chat Control 1.0 actually does, which has not changed materially since 2021, and more about what it signals.

A measure that was rejected by the EU Parliament in March, by a procedure where parliament’s own majority doesn’t bind the outcome, was back in force before the summer recess. The mechanism used has no clear precedent in EU legislative history.

If this weren’t worrying enough, the real battle lies ahead: Chat Control 2.0. While trilogue round six is expected to resume in September under the Irish presidency, the question it poses, of whether mass scanning of private communications can be made compatible with the EU Charter, and whether “voluntary” obligations can be agreed that will achieve detection without undermining end-to-end encryption, won’t be resolved this month or later in 2026. Most likely, such questions will be answered in a court, not a meeting room.

Indeed, a legal challenge before the Court of Justice of the EU is now significantly more likely, given that the Council’s own Legal Service has placed on record its concerns about the regulation’s compatibility with Article 7 of the Charter.

In the meantime, the practical steps for Internet users to take are the same as always: moving sensitive communications to services that are technically incapable of cooperating with a scanning order, even if they wanted to, as they protect all sensitive data with end-to-end encryption.

FAQ

What is Chat Control?

Chat Control is the nickname for two separate EU laws about scanning private messages for child sexual abuse material. Chat Control 1.0 is Regulation (EU) 2021/1232, a temporary exception to the ePrivacy Directive that permits providers to scan voluntarily. Chat Control 2.0 is the proposed Child Sexual Abuse Regulation, which would make detection mandatory. Only 1.0 is currently law.

Did Chat Control pass on 9 July 2026?

Yes. The European Parliament failed to block the Council’s second-reading position, so Chat Control 1.0 was reinstated and now runs until 3 April 2028. It had legally expired on 3 April 2026 after Parliament rejected an extension in March.

Why did it pass if more MEPs voted against it?

Because 314 votes to reject were not enough. Under second-reading rules, Parliament can only reject or amend a Council position with an absolute majority of all members, currently 361 of 720. Only 314 MEPs voted to reject, against 276 who voted to keep it. Abstentions and empty seats functionally counted as support for the Council.

What is the difference between Chat Control 1.0 and Chat Control 2.0?

Chat Control 1.0 permits scanning. Chat Control 2.0 would require it. 1.0 covers known material on unencrypted services through hash matching. 2.0 would allow detection orders, cover previously unknown material and grooming, and puts client-side scanning of encrypted services on the table. 2.0 is still stuck in trilogue after five failed rounds.

Are my WhatsApp or Signal messages being scanned?

No, not under Chat Control 1.0. End-to-end encrypted message content cannot be read by the provider in transit, so it cannot be scanned server-side. Parliament also adopted an amendment explicitly excluding E2EE services from the derogation’s scope, though the Council still has to accept or reject it.

Which services actually scan messages under this law?

The services known to scan are unencrypted or server-side encrypted platforms: Gmail, Facebook Messenger, Instagram direct messages, Skype, Snapchat, iCloud Mail and Xbox. Providers choose to participate. The law creates the legal basis, not the obligation.

Are my cloud backups scanned?

Potentially, yes, and this is the gap most people miss. A message can be end-to-end encrypted in transit and then written in plaintext to an iCloud or Google backup, which sits outside the encryption envelope. Turning off cloud backup of message history, or encrypting it locally first, closes this gap.

Does scanning happen without a warrant?

Yes. Chat Control 1.0 requires no warrant, no judicial authorisation and no prior suspicion. This is the core objection raised by MEPs, and by the Council’s own Legal Service, which reportedly warned in June 2026 that generalised scanning may be incompatible with Article 7 of the EU Charter of Fundamental Rights.

Does a VPN protect me from Chat Control?

No. A VPN encrypts the connection between your device and the internet. Chat Control scanning happens on the platform’s servers, or under Chat Control 2.0 potentially on your own device, and a VPN reaches neither of those places. Any VPN provider claiming otherwise this week is selling you something.

How accurate is the scanning?

Contested, and the official numbers are not reassuring. Germany’s BKA reports that roughly 48 percent of Chat Control alerts are not criminally relevant. Swiss federal police put the figure of machine-generated reports without merit at around 80 percent. The Commission’s own implementation report found false positive rates of up to 20 percent for AI classifiers assessing unknown material.

What happens if I am falsely flagged?

A false positive generates a report to law enforcement. Depending on jurisdiction, that can lead to questioning, a house search and device seizure. Holiday photographs of children at a beach are the example cited most often by police themselves. Around 40 percent of investigations triggered by these flags reportedly target minors rather than offenders.

Does Chat Control apply to me if I live outside the EU?

Indirectly, yes, if you communicate with people in the EU or use services operating there. Scanning is applied to communications, not to passports. A message from a non-EU user arriving at an EU-facing service can be scanned like any other.

We love bringing you this content and hope it helps keep you safe and secure online. Feel free to share it with your friends, too.

Here at hhgttyu.pages.dev we are all about internet freedom, and we are happy to be in a position to bring that to everyone. That is why we give you a 30-day money-back guarantee on our Premium plan. No questions asked and no logs recorded.

Get hhgttyu.pages.dev VPN

If you have any questions, please feel to contact our 24/7 support team either at support@hhgttyu.pages.dev or via live chat.

Related Articles

What personal data do age verification services collect and keep?
What personal data do age verification services collect and keep?

“A VPN can change the IP address a verification service sees, but…

Low price to sign up, higher price to renew: a common VPN industry practice
Low price to sign up, higher price to renew: a common VPN industry practice

“You signed up two years ago at a low monthly rate. Two…

Is the UK banning VPNs? What’s actually happening in 2026
Is the UK banning VPNs? What’s actually happening in 2026

“VPNs are fully legal in the UK today. But a targeted attempt…

D-Wave quantum annealing factoring RSA: Is it a real cybersecurity threat?
D-Wave quantum annealing factoring RSA: Is it a real cybersecurity threat?

In late 2024, headlines began to circulate in cybersecurity circles with titles…

Harvest now, decrypt later: Why attackers are stealing your encrypted data today
Harvest now, decrypt later: Why attackers are stealing your encrypted data today

Imagine you’re sitting in the lobby of a hotel in Tokyo. You’ve…

Age verification laws around the world: A country-by-country guide
Age verification laws around the world: A country-by-country guide

In the last twelve months, a wave of extremely stringent age-verification restrictions…

Freedom.gov: Free speech tool or government surveillance honeypot?
Freedom.gov: Free speech tool or government surveillance honeypot?

Freedom.gov is an as-yet unlaunched US State Department-developed portal, with planned built-in…

Your ‘free’ VPN is selling your ChatGPT chats, and you agreed to it
Your ‘free’ VPN is selling your ChatGPT chats, and you agreed to it

Imagine a situation where an ordinary, privacy-conscious person decides to install a…

Moltbook privacy risks: Can the AI-agent social network expose your data?
Moltbook privacy risks: Can the AI-agent social network expose your data?

In late January 2026, entrepreneur Matt Schlicht launched Moltbook, the very first…