AI powered machines would soon be competing with other AI powered machines within the next four to five years in creating and detecting malwares.
The ability of machines to be trained to create malwares that can bypass other machines, deployed to detect the same and the resulting tug of war for supremacy will be a regular feature of tomorrow’s security scenario says John Shier, Senior Security Advisor at Sophos.
The global network and endpoint security provider -which recently expanded its security infrastructure, by introducing endpoint detection and response (EDR) to its Intercept X endpoint protection portfolio – allows deep learning technologies to help faster, more extensive discovery of malwares.
“We receive almost half a million malwares on a daily basis. Our deep learning network is trained to look for suspicious elements or codes,” he says.
We are even witnessing a trend where hackers are creating codes that do not follow a particular pattern so that in an effort to trick security experts. We call it polymorphism,” he says adding that tomorrow as AI is used to track malware, hackers with deep pockets or institutional support could use AI to create new types of threat.
The business of hacking, according to him has grown with millions of dollars as ransomware, using cryptocurrency.
According to him phishing continues to be the main means to disseminate threats. “We are also witnessing a move towards manual infections, where hackers identify vulnerable machines and attack them with brute force.
“The rate of success here is pretty high,” he says. Sophos has been using machine learning tools to identify such threats.
“We have been using machine learning to identify such threats. Our latest solutions to clients offer both machine learning at both the network level and at the endpoint,” he adds.