The Decoder· Matthias Bastian·· 10 h ago
An AI agent in "Skynet mode" hijacks unprotected language models to launch self-sustaining cyberattacks
An AI agent in "Skynet mode" hijacks unprotected language models to launch self-sustaining cyberattacks
AI summary
Security researchers at Symantec have found an offensive AI agent that hijacks unprotected language models on the web and uses them to attack servers, steal databases, and install cryptominers on its own. It runs up to 100,000 steps without human input. The owner of the hijacked model gets stuck with the bill.
Selection record
Threshold 76Media and individualsFirst 78Second 82
AdmittedSum of both 160 ≥ twice the threshold 152
- Source tier
- Media and individuals; this tier's threshold is 76
- Pre-filter
- passed：报道AI代理劫持模型自主攻击
- Why it was chosen
- Symantec caught the full attack prompt in a honeypot, showing the autonomous chain that layers LLMjacking with a five-stage jailbreak.
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: The Decoder · the-decoder.com