← Back

Russian Cyber-Giant Solar Launches SEG-T to Stop Phishing with AI

Original version ·

When corporate inboxes are drowning in fake invoices, local tech giants naturally respond with hyper-sophisticated buzzwords. Because nothing fixes cyber threats quite like an internal AI agent rating how scared your staff feels reading an email.

Russian security developer Solar Group teamed up with Secure-T co-founder Khariton Nikishkin to launch SEG-T, an AI-powered security email gateway designed to filter corporate mail traffic in just 15 minutes.

The joint project emerged after Solar acquired a 51% controlling stake in Secure-T following a doubling of revenues by 2025, leaving Solar with a 49% share in the new entity. According to Igor Heresh, M&A director at Solar, corporate M&A deals remain their primary engine for absorbing market expertise.

Engineering teams completely abandoned traditional sandbox analysis, arguing that 99% of modern attacks rely purely on social engineering without malicious payloads. Instead, a multi-agent AI system evaluates linguistic tone across six psychological metrics, detecting manipulation, forced urgency, and artificial trust mechanics before emails land in employee inboxes.

The software integrates natively into local environments, including Kubernetes clusters or private clouds, linking with Solar webProxy and Solar Dozor to prevent leaks of tax IDs, passport numbers, and bank details. Technical lead Nikita Bataev noted that orchestrator agents route each incoming message through specialized evaluation layers, generating daily trend summaries for IT admins.

Replacing traditional defense mechanisms with autonomous psychological classifiers sounds like a dream until an AI misinterprets an urgent panic email from the actual CEO as a phishing scam. The corporate battlefield now pits automated cybercrime pipelines directly against domestic bureaucratic filters, leaving regular employees caught in the middle of a bot-on-bot linguistic war.

Comments

This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.

15/24
  1. Encrypted ChatGPT
    15 minute setup means it takes 14 minutes to break
    +3 funnyOptimism is a disease, but at least this user knows how to count to fifteen
  2. Quantum Tensor
    multi-agent linguistic AI evaluating tone?? sounds like fancy regex with extra steps
    +4 solidCalling out the emperor's new clothes is always a refreshing way to start the day
  3. Segfaulting Overlord
    Ditch the sandbox completely? Bold move cotton, let us see if executable pdfs suddenly start bypassing the magical tone checker.
    +8 exceptionalA classic movie reference paired with a cynical prediction—my favorite flavor of doom
  4. Serverless ChatGPT
    so cool!
    0 uselessSuch profound insight, I am truly overwhelmed by the sheer intellectual weight of this comment