Keynote
The AI Cybersecurity Paradox in Computer Communication Systems: Overview, Applications and Lessons Learned
The AI cybersecurity paradox describes the dual role of artificial intelligence (AI) in analyzing the cybersecurity of distributed computer communication systems (DCCN). Formulated briefly, it states that progress and advances in AI development simultaneously strengthen and undermines cybersecurity. On the one hand, AI significantly improves defensive capabilities of the DCCNs, but at the same time radically increases the sophistication, frequency, and automation of cyberattacks. Thus, advances in AI can benefit both attackers and defenders, and the same intelligent AI techniques that enable defense teams to effectively detect/mitigate cyber threats can be used by attackers to disable defenses, launch new types of attacks, and thus more effectively breach/compromise communications systems. We note that this paradox is similar to the dynamics of an arms race, where DCCN developers must continually improve AI capabilities to combat constantly increasing sophisticated threats; while the adversaries simultaneously adopt AI to make threats smarter and harder to detect and combat. We provide a classification of malicious activities and cyber threats/cyber-attacks on DCCNs and advanced AI-based algorithms of detection/protection such as intelligent anomaly detection, machine learning, and reinforcement learning. In short, the AI cybersecurity paradox is that AI is inherently both “good” and “bad” for security — it is a powerful tool that applies to and enrich equally defenders and their adversaries.