We know that AI is making a mark on the future of cybersecurity, as it's already a part of a substantial number of cybersecurity outfits. But to know where AI is headed as a cybersecurity tool, we have to monitor the trends impacting the technology. These industry insiders shared their views of the trends most shaping AI in cybersecurity. Here's what they said:
1. Stacy Stubblefield, co-founder and Chief Innovation Officer at TeleSign
“A trend we are starting to see more of is machine learning and behavioral analytics to identify fraudsters before the fraud attempt even occurs. The computer’s ability to analyze behavior trends is leading to filtering of payment fraud, fake reviews, fake comments and fake news being posted on some of the most popular online sites. With the power of machine learning in cybersecurity, we can even develop scores to rate specific users on their reputation and probability for fraudulent activity. As this year winds down, we’ll see these tactics with AI begin to further mature.”
2. Chris Day, Chief Cybersecurity Officer for Cyxtera
“We see the #1 trend is that vendors can’t simply use the hype surrounding AI in their products, but must start showing efficacy in enterprise operational environments. There is a lot of snake oil out there right now in this area.”
3. J.J. Guy, COO of JASK
“Alert fatigue and false-positive alerts continue to be issues leading companies to look to automation to up-level their security operations. While skilled SOC analysts are the very best defense we have in combatting threats, they’ll simply get burnt out if they have to manually look into each alert that crosses their desks.
Ideally, human SOC analysts should be focused on more advanced and more valuable tasks like threat hunting, threat intelligence, and attribution – the things they signed up for when they chose a career in security operations. By employing automation, the idea is to free analysts up to work on more important tasks and let automation do grunt work.”
4. Raul Popa, CEO of TypingDNA
“2019 is a year with significant AI improvements in fields like adversarial networks and transfer learning and we started seeing smarter apps that take advantage of these relatively new methods. Most AI applications for cybersecurity are focusing on applications of computer vision, and basic detection of anomalies like fraud or intruders, but also risk analysis. We see AI being employed everywhere to improve the accuracy of statistical models and predictions in detecting attacks of all kinds.
Andrew Ng in a recent interview said the #1 trend to look for in the next year(s) is “small data learning” – a new set of algorithms and methods allowing new players in the industry to take advantage of limited number of samples to yield predictions that, not a long time ago, were reserved only for a few companies that had big data to play with. Small data analysis also allows a lot of new problems to be tackled, including the ones we are addressing at TypingDNA, where we are able to differentiate an intruder from a genuine after training on very few previous samples (typing patterns in this case).”
5. Joshua Crumbaugh, Chief Hacker/CEO at PeopleSec
“The number one trend shaping AI this year is the lack of innovation. Very few companies are using AI to address the elephant in the room that is the human aspect of cybersecurity. Nothing we do will work until we address this problem and everyone knows it. I was able to figure out how to use AI to stop people from clicking on phishes and I believe that we are just scratching the surface of what is possible through AI.”
6. Dr. Murat Kantarcioglu, Professor of Computer Science at The University of Texas at Dallas
“I believe that many people in the industry have started recognizing that applying AI in cybersecurity is quite different compared to many domains. First of all, the attacker may deliberately poison the data collection process. Therefore, you may not have good enough data to train and build your AI models.
Even if you build the AI model, an attacker may manipulate the data under its control (e.g., modifying the spam email to bypass any spam filter even if the filter is based on advanced AI). Therefore, the realization that more robust techniques that are resistant to such manipulation is an important emerging trend.”
7. Carl Hasselskog, co-founder and CEO fo Degoo
“AI is bringing new functionality to operations that in the past have not been inherently user-friendly or intuitive. In the world of cloud storage, platforms are becoming more experiential for users, now more valuable to them in ways beyond their conventional role of purely existing for storage purposes. AI is capable of (securely) utilizing user data to provide a more personalized touch to various cloud functions whether it be saving information or organizing and reviewing files and photos.”
8. Aby Varghese, Chief Technology Officer at UIB
“The #1 trend shaping AI in cybersecurity this year are Blockchain technologies. IoT and AI require systems to be exposed to the internet — e.g., controlling connected devices with WhatsApp sends your information around the world, requiring secure data. This makes access to and cybersecurity of the underlying systems of prime importance. Companies are adopting Blockchain technologies to their IT systems to both prevent tampering and to immediately alert (and even take informed action) when tampering occurs.”
9. Kevin Landt, VP of Product Management at Cygilant
“One of the biggest trends is the replacement of signature-based threat detection with AI-driven detection. Instead of developing a signature for known malware after it's already been unleashed, threat detection platforms are starting to use AI to recognize anomalous behavior. This means you can detect malware and attack methods that researchers haven't even seen before.”
10. Eyal Benishti, founder and CEO of IRONSCALES
“Organizations and MSPs/MDRs are increasingly turning to AI-powered mail security orchestration, automation and response (M-SOAR) solutions to address the growing security threats.”
11. Einaras von Gravrock, CEO of Cujo AI
“The topic of data privacy is front and center in most of the discussion led by cybersecurity experts today. This is an exciting development, considering that just a couple of years ago, digital privacy was not seen as a priority by many. Today governments and business leaders, nongovernmental organizations, and local communities have started working together toward a global privacy awareness. This will hopefully turn into a comprehensive strategy that secures the right to digital privacy for everyone.”
12. Omar Yaacoubi, co-founder and CEO of Barac
“Undoubtedly one of the major trends in cybersecurity today is the rapid increase of encryption and the many problems it creates. As more organizations turn to encryption to protect their sensitive data – Google estimates that 80% of all traffic will be encrypted by the end of 2019 – and comply to more stringent privacy regulations, cybersecurity solution providers must look at ways to adapt their offerings to keep up.
Where AI comes in is that it can start to overcome some of the hurdles that crop up thanks to encryption, such as the amount of computing power it takes to decrypt and re-encrypt traffic to search for malware, not to mention the regulations the decryption method could be breaking in decrypting data. AI and its counterparts can offer new ways to look inside encrypted traffic, without breaking regulations or causing a detrimental effect on network performance.”
13. David Chavez, Vice President of Avaya Incubator at Avaya
“Complexity of modern Advanced Persistent Threats (APT) attacks, pivoting and distributed advanced evasion techniques (AET) evade human real-time pattern recognition. AI can assist recognizing, isolating, compensating and blocking potential threats. There is less of a need for hard-coded pattern recognition, but rather system-wide pattern tracking and moreover predicting disinformation and misinformation campaigns, war of cognition and similar.”
14. Anuj Goel, CEO and co-founder of Cyware
“-AI is a learning technology as more focus and effort are spent here the technology will only get better.
-Identification of adversary behavior and malicious signatures in this ever-changing threat landscape has emerged as a huge challenge for organizations that are always on their toes to continuously block threats or recalibrate their security tools against known threats. The trend shows that organizations are increasingly leveraging AI to identify both known and unknown threats and stop them.”
Have expert insights to add to this article?
Share your feedback and we'll consider adding it to the piece!ADD YOUR INSIGHTS