Telegram, Terrorists and the Right to Be an Idiot

On Wednesday, the Russian FSB charged the Telegram founder Pavel Durov with facilitating terrorism, accused the platform of failing to remove channels, chats and bots allegedly used by extremist organisations, and announced plans to put him on an international wanted list.

One allegation concerns a Telegram dating bot that Russian authorities say was used to recruit young people for sabotage.

Russia is not even first in the queue.

France arrested Durov in August 2024 and placed him under formal investigation over a collection of allegations concerning criminal activity on Telegram, including organised fraud, drug trafficking, child sexual abuse material and the refusal to provide information requested by authorities for legally authorised interceptions.

Telegram itself accepts that such material needs to be removed.

According to its current transparency figures, the company has blocked more than 21 million groups and channels during 2026, including more than 150,000 communities connected with terrorism. Telegram does not operate on the principle that criminal content must remain online in defence of liberty.

Tools like Telegram or Signal are used by normal people, conspiracy nuts, families, friends, casuals.

And yes, also criminals.

Thus, there is a perfectly legitimate question to ask here.

Does this mean that the state has the right to read all the communication, flag it and investigate?

No.

Private communication should be genuinely private, while criminal activity should still be investigated as criminal activity. Those are not contradictory ideas.

The state has legitimate reasons to investigate terrorism, trafficking, organised crime and recruitment networks. What does not follow is that every citizen's private correspondence should become machine-readable to make the investigator's job easier. That is essentially turning everyone into a permanent suspect.

Investigators should generally attack everything around the encryption.

Seize the unlocked phone. Compromise an endpoint. Infiltrate the group. Arrest one member and obtain his device. Exploit an implementation bug. Examine metadata. Follow the money. Run undercover accounts. Correlate identities across services. Watch public channels and then follow people into private networks. Get one participant to cooperate.

That is exactly how it should work and requires investigation.

The alternative being proposed repeatedly in Europe and elsewhere is essentially an attempt to replace investigation with infrastructure. Automated mass inspection is seductively scalable.

Suppose Signal implemented mandatory automated inspection tomorrow. Serious criminals would not throw up their hands and become accountants. They would move.

They can use another service, their own server, encrypted email, XMPP, Matrix, custom software, dead drops, game chat, shared cloud documents, steganography, code words, or a perfectly ordinary encrypted blob transferred through some unrelated service. Two competent people can create private encrypted communication without Signal, Telegram or WhatsApp existing at all.

That is the fundamental flaw in mass surveillance arguments.

You can regulate platforms. You cannot regulate mathematics.

The people most affected by weakening mainstream encryption are therefore disproportionately ordinary users, businesses, journalists, lawyers, dissidents and badly organised criminals. Sophisticated criminal organisations adapt.

“If somebody wants to spend six hours reading that Klaus Schwab communicates with reptilians through the CERN particle accelerator, or that the WHO is run by lizard people who meet in a volcano under Davos, I don’t particularly want Brussels rescuing them from themselves. Freedom includes the right to be an idiot.

Keep private conversations private. Investigate suspected private criminal activity through targeted police work. Accept that some criminals will successfully hide things because eliminating that possibility would require policing everybody.

A free society necessarily tolerates some capability for wrongdoing because every mechanism capable of preventing all private wrongdoing also prevents meaningful privacy.