Cyber criminals are getting smarter, which means enterprises and carriers must fight back smarter than ever before.
This was the main discussion during the NetEvents Global Press and Analyst Summit in San Jose, California.
The debate - "The smartest tool in the cyber security toolbox: artificial intelligence" - was facilitated by Mike Spanbauer, VP of research at NSS Labs.
The panellists were Ron Green, executive VP and chief security officer (CISO) at MasterCard; Gary Sockrider, principal security technologist at Abor Networks; Anup Ghosh, chief strategist for next-gen endpoints at Sophos; Oliver Tavakoli, chief technology officer at Vectra; Roark Pollock, senior VP of marketing at Ziften; and John Michelsen, chief product officer at Zimperium.
The panel discussed the real state of AI in security software and solutions, and talked about the strength and weaknesses of lab-based approaches to the ever-increasing source of malware and hack attacks.
Artificial intelligence (AI) is key, according to analysts and vendors in the market who are using advanced heuristics, machine learning, pattern recognition, big data, neural networks, fuzzy logic and data science to tackle some of the biggest challenges in security.
For example, the speakers noted, AI-based anti-malware solutions use millions of learnt patterns instead of virus signatures to detect documents and executionables that are malicious even if that particular malware file has never been seen before.
"Companies are being dragged to use AI and machine learning because it studies patterns of cyber crime using deep learning," said Ghosh. He believes there will be rapid adoption of AI and machine learning to combat cyber crime in the near future.
The panellists also talked about AI-based analytics being able to review and filter security alerts to reduce false positives, allowing security professionals and CISOs to focus on real challenges and opportunities affecting their organisation.
"Connected machines already outnumber people currently. So if we are to keep up, we have to make use of machine learning and AI," said Sockrider. For him, human error is the weakest link in regards to cyber crime, and AI can fill that gap.
AI-based endpoint and parameter security can fight varieties of real and potential breaches, while consuming fewer resources than traditional firewalls, anti-virus and intrusion or prevention systems, he concluded.
Share