Artificial intelligence (AI) in the right hands has become a tool to increase productivity, detect cancer, predict the weather, find protein configurations and, most recently, find security vulnerabilities in software and computer systems. Cloud security analysts Sysdig reported earlier this month that they had discovered JadePuffer, the first ransomware controlled by an AI agent.
JadePuffer does the job of attacking systems without human guidance. Instead, the AI agent will carry out the attack process step by step until the attacked system is successfully penetrated. It can adapt the attack to human thinking. For example, it can retry failed steps within refined parameters.
After the data is accessed, it is exported before the original files on the attacked system are deleted. JadePuffer then leaves a note that the system has been attacked and provides the ransom amount to be paid, a Bitcoin account and an email to be automatically contacted. AI agents that were developed to perform repetitive tasks for humans in the workplace are now entering the era of hackers.
