Telegram used to be the quiet corner of the internet where journalists could breathe. No feeds. No trending lists. No invisible algorithms deciding what you see. Just channels-organized, chronological, and transparent. If you wanted to find a source on Ukraine grain thefts, protest footage from Belarus, or leaked documents from a government agency, you searched. You subscribed. You scrolled. And you knew exactly what you were looking at.
That changed in late 2023. Telegram rolled out a recommendation engine-not the kind that tracks every like, share, and scroll like TikTok or YouTube, but something quieter, more limited. It suggests channels based only on the topics of the channels you already follow. No engagement metrics. No behavioral profiling. Just: "You follow news about Russia? Here are other news channels about Russia."
At first glance, it sounds harmless. Maybe even helpful. But for journalism, especially investigative journalism, this small shift carries big consequences.
Why Telegram Was the Journalist’s Secret Weapon
Before recommendations, Telegram was unique because it didn’t pretend to be social media. It didn’t need to. Journalists didn’t use it to go viral. They used it to find truth.
Here’s how it worked:
- Channels posted in reverse chronological order. Every message was visible. Nothing got buried.
- File sharing went up to 2GB. A reporter could drop a 1.8GB video of a protest or a 500-page PDF of financial records without compression.
- There were no likes, shares, or comments. No way for bad actors to game the system by buying bots or spamming hashtags.
- Secret Chats gave reporters and sources end-to-end encryption. No cloud backups. No metadata leaks.
- Users could mute or unmute channels at will. No algorithm forced you to see content you didn’t want.
Tools like Telepathy were built by investigators to fill the gaps. They could map how channels linked to each other, trace the origin of images across dozens of channels, or pinpoint where a protest was being discussed based on geotagged messages. All of this was possible because Telegram didn’t mediate the flow of information. It just delivered it.
That’s why, in 2020, during the Belarus protests, Telegram became the most trusted news source for demonstrators. Not because it was flashy. But because it was reliable.
The New Recommendation Engine: What Changed
Telegram’s recommendation system doesn’t analyze how long you watch a video or whether you reply to a post. It doesn’t track your clicks. It doesn’t even know if you read the messages in a channel you’re subscribed to.
It only looks at one thing: the topic of the channels you follow.
So if you follow a channel about Ukrainian defense updates, Telegram might suggest another channel about Russian military logistics. If you follow a fact-checking bot, you might get recommended another fact-checking channel. Simple. Transparent. Limited.
That’s a big difference from Instagram, which processes 65 billion data points every second to decide what you see. Or TikTok, which scans video descriptions, audio trends, and facial expressions to predict what will keep you scrolling.
Telegram’s system is like a librarian who knows you like history books and says, "Here are some other history books." Not like a salesman who knows you bought a book last Tuesday and now keeps pushing you five similar ones-even if you never opened them.
But here’s the catch: even a simple recommendation engine changes how people discover information.
How This Affects Investigative Journalism
For casual users, recommendations are a convenience. For journalists, they’re a potential blind spot.
Before, if you wanted to track a disinformation network, you had to dig. You’d follow a channel that posted leaked documents. Then you’d find another channel that shared the same image. Then you’d trace it back to a third channel that originated the post. You built the map yourself.
Now, Telegram might suggest a channel you didn’t know existed. But why was it suggested? Because it shares keywords with a channel you follow? Because it’s popular in a certain region? Because it’s run by someone who’s also running five other similar channels?
You don’t know. And that’s the problem.
Investigative journalism thrives on transparency. You need to understand the source, the chain of custody, the network. If an algorithm quietly introduces you to a channel that looks legitimate but is actually part of a coordinated influence operation, you might miss the red flags-because the platform already vouched for it.
And that’s exactly what happened on Twitter in 2014. At first, it was a great place to find breaking news. Then the algorithm started favoring outrage. Then misinformation spread faster than facts. Journalists lost control of their own discovery process.
Telegram is still far from that point. But it’s taking the first step.
What Journalists Can Still Rely On
The good news? The core of Telegram hasn’t changed.
Every channel still posts chronologically. Every message is still there. You can still search by keyword. You can still subscribe directly. You can still use Telepathy and other tools to map networks, reverse-search images, and track geolocations.
The recommendation engine is just a sidebar. It doesn’t replace search. It doesn’t override your feed. It doesn’t hide posts. It’s an optional tool-like a suggested reading list at the bottom of a library catalog.
Journalists who understand this can still use Telegram the way they always have. They just need to be aware that newcomers might not.
Young reporters, freelance investigators, or citizen journalists who aren’t trained in open-source research might rely too heavily on recommendations. They might assume the algorithm is filtering for quality. It’s not. It’s filtering for topic similarity. A channel could be full of misinformation, but if it uses the same keywords as a trusted source, it’ll show up.
The Bigger Picture: Is Telegram Becoming Like the Others?
Telegram’s founders have always said they don’t want to be social media. They’ve resisted ads. They’ve refused to sell user data. They’ve avoided engagement-based algorithms.
The recommendation engine feels like a compromise. Not a betrayal. But a compromise.
It’s not trying to keep you addicted. It’s trying to make the platform easier to use. That’s not evil. But it’s not neutral either.
Every time a platform adds algorithmic discovery, even a mild one, it shifts power. The power to surface information moves from the user to the platform. And once that power is handed over, it’s hard to take back.
Telegram still gives users control. But the door is open. And once it’s open, others will try to push it wider.
What Journalists Should Do Now
Here’s how to adapt:
- Use recommendations as a starting point, not a destination. If Telegram suggests a channel, verify it. Check its history. See who else is following it. Look for patterns.
- Keep using search. Don’t rely on suggestions. Type in keywords. Use Telegram’s built-in search tool. Find channels yourself.
- Stick with trusted tools. Telepathy, InVID, and other open-source investigation tools still matter. Recommendations don’t replace them.
- Teach others. If you’re mentoring new journalists, don’t just show them the recommended channels. Show them how to dig deeper.
- Watch for changes. If Telegram starts using engagement signals-likes, shares, replies-then it’s time to reconsider your reliance on the platform.
Telegram isn’t broken. But it’s changing. And journalism has to change with it.
What’s Next for Telegram Journalism?
The future isn’t clear. Will Telegram keep its recommendation system simple? Or will pressure from advertisers, governments, or user growth push it toward more invasive algorithms?
Right now, it’s still the best platform for open-source journalism. It’s faster than Twitter. More secure than WhatsApp. More transparent than Facebook. And it still lets you share files larger than most email attachments.
But the moment you start trusting an algorithm to find your sources, you’ve already lost a piece of your control.
Journalists have always been hunters. They don’t wait for food to be served. They track it down. Telegram’s recommendation engine might hand them a map. But they still need to know how to read it.
Is Telegram still safe for journalists to use after the recommendation engine?
Yes, Telegram remains one of the safest platforms for journalists. The recommendation engine doesn’t compromise encryption, file sharing, or channel transparency. Secret Chats still work the same. Cloud Chats are still unbacked up by default. The algorithm only suggests channels based on topic, not behavior. Journalists can still use Telegram exactly as before-they just need to be aware that new users might rely on recommendations without verifying sources.
Do Telegram recommendations boost misinformation like other platforms?
Not yet. Unlike TikTok or YouTube, Telegram doesn’t use engagement metrics to promote content. It doesn’t know if you watched a video, liked a post, or spent time reading a message. It only looks at the topic of channels you follow. That makes it much harder for misinformation to go viral through the recommendation system. However, if a misleading channel shares keywords with a trusted one, it could still appear in recommendations-so verification remains critical.
Should I stop using Telegram’s recommended channels?
No-but treat them like a suggestion, not a guarantee. Use them to find new channels, then investigate them the same way you always have: check their posting history, see who else follows them, look for signs of coordination or manipulation. The algorithm doesn’t verify content. Only you can.
Can I turn off Telegram’s recommendation engine?
You can’t disable it entirely, but you can ignore it. Recommendations appear in a separate section called "Recommended Channels." They don’t appear in your main feed. You can still browse channels directly, search by keyword, or subscribe manually. The engine is optional. It doesn’t replace your control over what you see.
What tools should journalists use alongside Telegram now?
Telepathy remains the most powerful tool for mapping Telegram networks. It lets you search messages across channels, reverse-search images, track geolocations, and visualize how channels are connected. Other tools like InVID and Bellingcat’s OSINT tools still apply. Telegram’s recommendations make discovery easier-but they don’t replace the need for deep analysis.
Telegram still holds the keys to open-source journalism. But the lock is changing. The key still fits. But now, someone else has a copy-and they’re handing it out to anyone who asks.
