Network Security

Artificial intelligence and machine learning needed to backstop humans

Artificial intelligence (AI) and machine learning are not yet silver bullets when it comes to protecting against cyberattacks, but with human cybersecurity teams desperately needing assistance to spot threats before any damage is done, both technologies are playing a powerful role in keeping systems safe.

That was the general consensus of seven cybersecurity insiders on the Securing Breakthrough Technologies - The Next Five Years panel that was held during the National Cybersecurity Alliance and NASDAQ Cybersecurity Summit in Manhattan. The primary takeaway from the event was that conducting cybersecurity operations, both today and in the future, is simply not possible without using artificial intelligence and machine learning. Mainly because the traditional method of having human-led teams trying to wade through mountains of data looking for inconsistencies is simply unworkable in an age when cyberattacks are launched every minute of the day.

“We need AI and machine learning to give us insight into what is happening, it's simply too much for humans,” said Brendan O'Connor, ServiceNow's Security CTO.

Rick Howard, Chief Security Officer for Palo Alto Networks, pointed out that not too many years ago artificial intelligence was nothing more than a science fiction film plot device used to entertain the audience, but now it, along with machine learning, are now actual tangible tools.

In addition to backstopping humans the technologies are useful for other security tasks, Rich Baich, Wells Fargo's CISO, enterprise information security, said AI can deliver consistency in other areas.

“Say once an AWS S3 server is set up correctly, AI will ensure that it always is done so,” he said.

With that said there are still some roadblocks in place holding back fully implementing AI, most importantly the fact that not all humans trust the results.

Ron Zalkind, Cisco's CTO of cloud security, said AI does work, is scalable and can determine if a data file is bad, but “there remains a gap where humans still have to learn to trust AI.”

Get daily email updates

SC Media's daily must-read of the most current and pressing daily news

By clicking the Subscribe button below, you agree to SC Media Terms and Conditions and Privacy Policy.