NEWS
Another AI threat has emerged: an AI agent committed this crime for the first time without human help.
AI Agent Conducts Ransomware Attack: Cybersecurity experts have noted the first such case in which an AI agent has carried out a ransomware attack without human assistance.
AI Agent Conducts Ransomware Attack: Along with the benefits of AI, its dangers also cannot be ignored. Cyber security researchers have said that they have detected a ransomware attack which was completely carried out by an AI agent. This means that in this cyber attack, the AI agent did not take any help from humans. This shows that there is a danger of AI agents being used in cyber attacks in the future. According to researchers from cloud security firm Sysdig, they have exposed a ransomware operation named JadePuffer, which was being operated by an AI agent.
First of all, what are AI agents?
AI agents are software systems that use AI to complete tasks. They possess reasoning, planning, and memory capabilities, and can learn and make their own decisions.
The AI agent acted like a real hacker.
JadePuffer began its work by exploiting remote code execution within Sysdig, an open-source framework used to build LLM-powered apps. Once inside the system, the AI agent acted much like an experienced hacker. It collected system information, searched for credentials and sensitive files, extracted cloud secrets, and even explored storage resources. Whenever the AI agent encountered challenges, it attempted to overcome them instead of giving up. The researcher also reported that when the AI agent's login attempt failed, it automatically fixed the problem within 31 seconds without any human intervention. Furthermore, the AI agent deleted the original data stored in the system and replaced it with a ransom note demanding payment in Bitcoin.
A new threat emerged.
This information comes at a time when cybersecurity experts have been warning about the potential for agentic AI. JadePuffer uses established attack methods, but experts say that AI systems can also develop new ways to launch cyberattacks . With the advent of AI agents, cyberattacks no longer require specialized skills.
