Areas of technology with large data sets and significant costs associated with response times offer unique opportunities for artificial intelligence solutions — and cybersecurity is no exception. In a new video, we sat down with Alex Chitea, a technology analyst who leads cybersecurity efforts at Orange Silicon Valley. He discussed the state of cybersecurity, how machine learning is currently used within the field, and different algorithms that detect cyber threats.
1. How can AI be used in cybersecurity?
“According to a report, financial institutions receive on average 200,000 alerts each day, which makes it impossible for human analysts to find the signal among this limitless noise. This is where artificial intelligence, and machine learning specifically — a subset of artificial intelligence — can step in and help.”
2. Which types of machine learning have specialized roles in cybersecurity?
“In cybersecurity, supervised algorithms is where machine learning has made the biggest impact. With supervised learning, an algorithm is essentially trained on past data threats and when trained well, it’s reasonably expected to catch those threats in the future. …Unsupervised learning looks at unlabeled data sets and tries to find the anomalous behavior, usually using very few data points to do so.”
3. What defines a competitive cybersecurity company?
“When I look at a company that claims to do some form of artificial intelligence or machine learning, I ideally look for three things. Number one, I look for unique data sets that are of good quality that the company owns exclusively, generates, or has access to. Number two, I look for expertise, either in the form of a proprietary algorithm that they have or the right team in place to produce IP. And last, I look [for] whether or not the company is solving an actual need — an actual problem.”