The Other Side of the Artificial Intelligence Cyber Coin
Chuck Brooks · 2024
Though useful tools for cyber defense, artificial intelligence (AI) and machine learning (ML) potentially have drawbacks. Threat actors can employ them to do their bidding. The cybersecurity landscape is quickly changing, encompassing AI-powered malware, adversarial ML techniques, AI-enabled botnets, and intensified insider threats. Cyber threats were growing globally faster than cyber defensive capabilities even before AI arose and was used. Cyberattacks in the past required much manual investigation, labour-intensive planning, and patience. Cybercriminals are also taking advantage of vulnerabilities in the ecosystem of the growing networks of connected devices. Cyberattacks will become more sophisticated as result of generative AI, increasing the possibility of ransomware-as-a-service, phishing, malware development, deepfakes, and the exposing of personally identifying information. Another kind of adversarial attack in the context of AI and ML models is AI poisoning. Polymorphic malware leverages polymorphism to evade detection instead of optimizing its effectiveness.