A neural network designed to identify owners of anonymous Telegram channels has emerged in Russia. The company behind the project, T.Hunter, describes the system as a sophisticated AI named “Comrade Major.” It has undergone internal testing, according to reports cited by NTI SafeNet, a market expert within the National Technology Initiative (NTI).
In discussions with the press, NTI SafeNet noted that the program is built on a foundational library of methods and techniques used by seasoned investigators when researching Telegram channels. Those techniques were digitized and then taught to the AI, creating a tool that can operate on the same principles an expert would follow. Igor Bederov, who leads the information and analytics research department at T.Hunter, emphasized in an interview that this approach mirrors the analytic workflow of a human investigator but is implemented in software form.
The described workflow involves gathering data from a channel’s description, archived posts, and the body of content published within the channel. The neural network examines a wide range of materials, including message threads, embedded chats, sticker sets, documents, and videos, to extract identifying information. The aim is to connect digital traces—such as usernames, photos, and patterns in behavior—to real-world identities. The developers claim the AI can perform these tasks more quickly and with greater consistency than a human analyst would be able to achieve in the same timeframe.
When prompted, the AI can present information related to mobile devices linked to a user profile, physical address data, and other digital footprints that may appear within the analyzed material. The system is also reported to scan additional identifiers and to pull data from associated communities on social networks, blogs, and various websites.
According to Bederov, the underlying technology runs on one of the most widely used text-based neural networks and operates from T.Hunter servers. A Telegram bot has been created to streamline interaction with the software, making it easier for investigators to initiate analyses and access results.
The developers also highlighted a second capability: the ability to verify the legal entities associated with a given tax identification number (TIN). This function adds a layer of due diligence for organizations that need to confirm corporate identities in investigations or compliance reviews.
Plans for a full release of the program were announced, with a timeline aimed at 2024 to 2025. The stated purpose of the application is to support public and private entities engaged in cybercrime investigations that involve anonymous Telegram communities.
There has been ongoing public concern about the use of Telegram bots and the potential for abuse. Russian observers have warned about the dangers that can arise when automated tools gain access to private and semi-private channels. In related discussions, questions have been raised about the balance between security and privacy, as well as the risks of misidentification when algorithms interpret online activity.
Historically, researchers have noted that Telegram’s platform can present opportunities and challenges for investigators. Earlier tools sometimes resembled automated “eyes” that could surface patterns across a user’s digital footprint. The emergence of a formalized AI system that aggregates data from multiple sources marks a notable step in the way authorities and organizations may approach cyber investigations and identity verification in the future.