Teleconsensus AI bot uses Hedera Consensus Service to create a verifiable and timestamped log of Telegram messages
Feb 13, 2020
by Teleconsensus

Telegram is the defacto social media platform for distributed ledger technologies (DLTs) and cryptocurrency communities to discuss their projects. It’s also a place for the founders, teams and developers behind those projects to provide updates to the community and to answer questions.

DLT platforms are all about providing an immutable source of truth, yet Telegram itself is far from being an immutable source of truth; Telegram includes the ability to delete a message or edit a message. While you can tell if a message has been edited in Telegram, you can’t tell what the original message said. This can create problems for both the teams behind the DLT projects and for the DLT project community members. Situations can arise where one is simply editing a typo in a Telegram message, but one can’t show if something was edited in retrospect. When a message is deleted there is no obvious way to determine whether the message was indeed deleted or what the deleted message said.

Many resort to taking screenshots of messages posted in order to protect themselves and to show what they have said or to capture a message that one believes is not true to hold as a record. The trouble with screenshots is they can easily be photoshopped and don’t necessarily prove when the screen shot was taken or in which Telegram group it was taken, therefore they are an unreliable source of evidence for proving what was said by who and when.

This creates a significant problem for DLT communities which need to have an accurate audit log of what was said by everyone in the community including the DLT platform team and founders as well as the community members.

These problems became apparent to several members of the Hedera Community and we decided to look towards Hedera Consensus Service as a way to provide a solution to this problem. This is where the brand Teleconsensus emerged: “tele-” is a Greek word often used as a prefix to mean at a distance and consensus being the term used to describe the process by which distributed computer networks agree on the order and state of things.

Teleconsensus intends to do many things in the future in the areas of DLT and Artificial Intelligence and the Telegram bot @teleconsensus_bot is the first project.

As it stands there are several Telegram recorder software which can operate on a desktop computer and record Telegram messages. These Telegram recorders can record any public or private Telegram group or private Telegram conversation potentially.

Telegram themselves say in their Safety and Security section of their privacy policy :

“To improve the security of your account, as well as to prevent spam, abuse, and other violations of our Terms of Service, we may collect metadata such as your IP address, devices and Telegram apps you've used, history of username changes, etc. If collected, this metadata can be kept for a maximum of 12 months.”

There are also existing Telegram recorder bots which are acceptable within Telegram and Telegram provides the following disclaimer within the terms of service:

“Bots added to groups can operate in two modes: with access to messages in the group or without access. If the bot has access to messages, it can see everything that happens in the group. The interface clearly shows whether or not a bot has access to messages in groups.”

Another option is to create a custom Telegram client and keep all logs of messages; to the other users of a group, this client would appear as a normal user.

Given these options to create a Telegram recorder we chose to use a Telegram bot to be fully transparent and to show that the bot has access to read messages, unlike other Telegram recorders which may currently be run by any member of the community without the community's knowledge.

However, recordings via desktop applications or by Telegram bots are no better than the screen shot solutions as messages recorded by these Telegram recorders can be edited or deleted or even forged creating the same problem as photoshopped Telegram messages.

So what's the right solution to this problem?

This is where the Trust Layer of the Internet, Hedera Hashgraph comes in to solve this problem in a way that could never have been solved until the advent of the Hedera Consensus Service.

How does the @teleconsensus_bot work ?

  1. Firstly, the @teleconcensus_bot needs to be invited to a public Telegram group and made an admin of the group with read privileges. As soon as the bot is added to a group it will begin to record any message that is posted to the group or any edit to an existing message.
  2. @teleconcensus_bot will send a copy of these recorded messages to the software on a secure server (Teleconsensus AI).
  3. A copy of the message is sent to a private Telegram group No. 1 including the Telegram update ID.
    1. At the same time, the software is sending a hash of the message along with the update ID to Hedera HCS. Note that a Hash is not like sending the full message in clear text to Hedera. Hedera does not see the original message, just the hash that is produced by that message.
  4. Once the Hedera Consensus Service puts a time stamp on the Hash of the message it comes back from Hedera along with the unique Telegram update ID.
  5. The Hash, the Telegram update ID, the Timestamp, the sequence number, and the running-hash are posted as a return message in private telegram group number 2.
  6. A clear text message in private group No 1 can then be matched with the HCS output that is held in private group No 2.
  7. In order to prove validity of a message, a user needs to request the clear text message, match it with the output of HCS and check the message signature to convince himself that the message was originated by HH indeed and that it was seen at the stated Timestamp.

The future of Teleconsensus Bot

We are also working on transparent logs of reputation and experience points, and audit logs of Telegram moderation actions.

If you would like to follow our progress and learn more about how to implement Hedera Consensus Service in your own project and be alerted when we make our code open source please follow us on Twitter at @teleconsensus or subscribe to our newsletter at teleconsensus.ai or join us on Telegram at t.me/teleconsensus.