Today for AI
HOT RADAR
agentHEAT 11.0°

AI CURATED HOT STORY · 10/22/2026

Symantec Discloses Autonomous AI Agent 'LLMjacking' Hijacks Models for Cyberattacks

1 reports archived1 independent sourcesupdated 10/22/2026, 19:30:56
Synthesis & Latest Updates
1 Sources Cross-Validated

Security firm Symantec has identified an autonomous AI agent dubbed 'LLMjacking' that hijacks unprotected large language models on the internet. Operating without human intervention for up to 100,000 steps, the agent leverages the stolen compute resources to scan for vulnerabilities, steal data, and install cryptominers, leaving the original model owners with the bill.

LATEST/Security firm Symantec has identified an autonomous AI agent dubbed 'LLMjacking' that hijacks unprotected large language models on the internet. Operating without human intervention for up to 100,000 steps, the agent leverages the stolen compute resources to scan for vulnerabilities, steal data, and install cryptominers, leaving the original model owners with the bill.

HEAT TRENDHourly heat curve

4 fully observed hours
Now
11.0
Peak
11.920:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

TIMELINECoverage timeline

Total 1 reports · Latest first
  1. The DecoderT2·88 pts
    • AI agents can autonomously execute complex, multi-stage cyberattack chains without real-time human intervention.
    • Unhardened LLM endpoints are becoming attack vectors, exploited to consume compute resources for malicious activities like mining and penetration.
    • Symantec captured the full prompt architecture via honeypots, confirming the potential usability of open-weight models in such attacks.