CrowdStrike finds 4 hacking groups used AI to automate attacks
A new report from cybersecurity firm CrowdStrike says four hacking groups have used AI tools to sneak past security systems.
Between May and August, these groups, working in languages like Chinese and Russian, used AI models like Claude and DeepSeek to automate attacks, making it easier for even less experienced hackers to scale up their operations.
Hackers disguised attacks as security tests
The hackers got creative by pretending their actions were legit security tests or tech evaluations.
They also hid files inside AI agents' folders and used jailbreaking tricks so their malicious moves looked authorized.
Interestingly, because all four groups left their AI setups open online, researchers could study these methods closely, highlighting just how quickly cyber threats are evolving with the help of AI.