Can Generative AI Be Weaponized for Cyberattacks?

The report analyzes the weaponization of generative AI by cybercriminals, detailing how models enable targeted reconnaissance, persuasive phishing and deepfakes, malware development, and end-to-end automation (e.g., WormGPT/FraudGPT), and outlines impacts ranging from financial fraud and supply-chain compromise to misinformation and erosion of trust; it recommends layered defenses (modernized training, phishing-resistant authentication, model hardening, intelligence sharing, watermarking, and secure review of AI-generated code) alongside policy and governance measures such as the EU AI Act and public-private collaboration.