In this special guest feature, Hal Lonas, Chief Technology Officer at Webroot, suggests that almost every software and information product and practice are thinking not “if,” but “when and how” to apply AI and ML. If they don’t, the competition will pass them up or make them irrelevant soon. There are many factors that have converged to enable this leap forward – let’s look at some of them. Hal Lonas is CTO at Webroot, a privately held internet security company that provides state-of-the-art, cloud-based software as a service (SaaS) solutions spanning threat intelligence, detection and remediation. Previously the Senior VP of Product Engineering for Webroot, Lonas has 25+ years of experience in enterprise software and engineering. He joined Webroot via the acquisition of BrightCloud, where he was a founder and VP Engineering. Lonas has held key engineering management positions with Websense (WBSN), ADP and others, and has co-authored several patents. He holds a Bachelor of Science in Aeronautics and Astronautics from MIT.
Artificial Intelligence (AI) and Machine Learning (ML) are getting a lot of press coverage these days and for good reason — it’s a broad and rapidly advancing technical frontier. Self-driving automobiles, commercial space flight, facial recognition, targeted ads, medical advances, and fake news detection all benefit from advances in AI.
Almost every software and information product and practice are thinking not “if,” but “when and how” to apply AI and ML. If they don’t, the competition will pass them up or make them irrelevant soon.
There are many factors that have converged to enable this leap forward – let’s look at some of them.
Data Gathering and Processing Power in the Cloud
We’re now able to gather and retain information in much more depth than ever before, because networking and storage costs have dropped. We also can keep this data in the cloud, which enables redundancy and processing power never before available.
These same trends that have reduced data costs have also brought more processing power to problems, and an ability to scale up and down dynamically. We can spin up a few cloud instances to experiment, and then scale them up to production when we achieve success.
Machine Learning Breakthroughs
By focusing on solving pragmatic problems, ML advances have produced steady-state improvements in many fields. Neural networks, deep learning, and algorithm improvements have all been brought to bear. For example, NASA’s planet-hunting Kepler space telescope recently detected an exoplanet using Google’s ML technologies.
Scarcity of Technologists / Huge Volume of Data
It’s great to have all this data, but what are we doing with it? We can’t afford to hire armies of technologists and analysts to find patterns and draw conclusions. However, AI and ML can leverage a small number of humans to mine the data and pull out relevant trends for further analysis.
Moving Down Market
Problems that used to be solely in the realm of the largest governments and enterprises are now addressable by individuals and small- and medium-sized businesses. These smaller entities have found they can leverage the new data, processing, and machine learning resources available to them, and have access to products and services that amplify their workforce. Look at history and think about computers becoming personal, and spaceflight becoming something where private citizens can innovate.
Walk then Run
While many are afraid of losing their jobs or relevance in this new world, we need to be realistic about what’s really possible today. In fact, we’re a long way from replacing humans in fields that need creativity and insight.
Science recently shared that AI needs masses of data and well-defined inputs and outputs to do well. AI can’t deliver on any task where a thorough, clear explanation is required. It uses stats and probability, so it never delivers a single optimal solution. Machines are less adaptable than us, so their tasks can’t change suddenly. It’s important to realize these limitations.
Leverage
Clear wins to date with AI and ML have been driven by amplifying human capability – for example a recent analysis by the US Defense Advanced Research Projects Agency (DARPA) showed that battlefield casualties rise when soldiers suffer from exhaustion and muscle failure – so they have started focusing on suits and devices that amplify and work with muscles to alleviate fatigue and stress.
This same approach is working today to identify security threats and remediate them automatically in the cyber IT battlefield. Machines can help with the menial chores of sifting through data, looking for patterns, and remediating simple attacks. Humans can be alerted when more sophisticated decisions and analysis are required.
Using machine-enhanced analysis and remediation, a group of two security analysts can suddenly do the work of twenty.
What’s Next
Despite the clear advantages, there will always be naysayers and those afraid of change. One thing that will change their minds is the rise of adversaries using AI and ML to enhance attacks. There is evidence that this is already happening with phishing pages that mimic their financial service or social website targets better than ever before.
Let’s be guided by history. Chess master Garry Kasparov who famously lost to IBM’s Deep Blue over 20 years ago said “human plus machine is better than either human or machine.” Those who figure out how to leverage machines and automation to work better and smarter will succeed. Those who don’t, will ultimately fail.
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind