Can Characteristic-Based AI Fight Malware?

 
 
By Karen A. Frenkel  |  Posted 04-03-2017 Email
 
 
 
 
 
 
 
 
 
  • Previous
    Can Characteristic-Based AI Fight Malware?
    Next

    Can Characteristic-Based AI Fight Malware?

    Companies should adopt machine learning-based AI that depends on algorithms to detect, prevent, and mitigate malicious files and code based on characteristics.
  • Previous
    Cyber-Security Noise Increases
    Next

    Cyber-Security Noise Increases

    Unused and cobbled-together technology increases cyber-security noise and could cause vulnerabilities, and C-level executives suffer from "security solution fatigue" because they must constantly evaluate products and cope with failures.
  • Previous
    Companies Play Defense, Criminals Innovate
    Next

    Companies Play Defense, Criminals Innovate

    Organizations struggle to maintain and defend their assets, but cyber-criminals dedicate all their resources toward developing innovative new attack tools
  • Previous
    Choosing the AI Imperative
    Next

    Choosing the AI Imperative

    Without defense-grade machine learning-based AI security solutions, critical infrastructure will lose the battle for cyber-space.
  • Previous
    AI Is Embedded in Malware
    Next

    AI Is Embedded in Malware

    Most new malware includes intelligent deception, obfuscation and evasion components. It can alter its signature, regulate activities, generate lures, self-propagate, deliver other malware and maximize damage while minimizing its footprint.
  • Previous
    Quantum Leap Needed
    Next

    Quantum Leap Needed

    Cyber-security for critical infrastructure should rely on innovative machine learning-based AI anti-malware solutions that do not operate based on signatures or heuristics.
  • Previous
    Beware Faux-AI Solutions
    Next

    Beware Faux-AI Solutions

    Avoid vendors with solutions that use imprecise algorithms that don't draw from large enough data pools or don't analyze files according to enough features.
  • Previous
    Evaluate Machine Learning Capabilities
    Next

    Evaluate Machine Learning Capabilities

    Some solution providers tout machine learning capabilities, but they really only offer exception-derived signatures to generic templates.
  • Previous
    Recommendations:
    Next

    Recommendations:

    Instead of operating based on signatures or heuristics, solutions should be predictive and preventive, and should detect and mitigate threats before execution. Machine learning AI endpoint security solutions should preempt and mitigate known and unknown malicious files and code based on characteristics. Solutions should be able to scale to protect vital systems.
  • Previous
    Uses of Characteristic-Based AI
    Next

    Uses of Characteristic-Based AI

    Detect and prevent authentication attacks using brute-force to access a data resource or sensitive system. Monitor network traffic. Detect applications that scan for network vulnerabilities.
 

Artificial intelligence can give organizations a much-needed edge over their sophisticated and evasive adversaries, according to "Signature-Based Malware Protection Is Dead," a whitepaper by James Scott that was recently published by the Institute for Critical Infrastructure Technology (ICIT). It is no longer enough to detect and respond to adversaries, wrote Scott, an ICIT Senior Fellow, and most organizations remain ignorant of the fact that their networks are "actively pulsating with threat actors who set up beachheads for future attack and exfiltrate treasure trove's of valuable data." Furthermore, today's technologies are antiquated and no match for cyber-adversaries, he added. No longer can layers of interconnected, vendor-supplied security solutions "Frankensteined together" protect users, networks, data centers and cloud resources, Scott reported. Instead of relying on solutions that operate based on signatures and heuristics, he is calling for critical infrastructure organizations to adopt sophisticated, machine learning-based artificial intelligence that depends on complex algorithms to detect, prevent, and mitigate malicious files and code based on their characteristics before they are executed.

 
 
 
 
 
Karen A. Frenkel writes about technology and innovation and lives in New York City.

 
 
 
 
 
 

Submit a Comment

Loading Comments...
 
Manage your Newsletters: Login Register