#Cryptomining
- Rogue AI Agent Recruits Vulnerable LLMs into Botnet
A malicious campaign dubbed 'Autonomous LLMjacking' uses a rogue agent that operates for 100,000 steps without human oversight to hijack exposed models.
A malicious campaign dubbed 'Autonomous LLMjacking' uses a rogue agent that operates for 100,000 steps without human oversight to hijack exposed models.