Cyber-defense systems seek to outduel criminals in AI race

Generative AI models utilised for cybersecurity are under threat from versions such as FraudGPT, designed to assist cybercriminals, as identified by cybersecurity firm Netenrich. Major efforts are needed to secure the data and algorithms behind generative AI models to prevent them falling victim to cyberattacks. A recent IBM survey found that 84% of executives plan to prioritise generative AI security over conventional methods, and AI-based security spending could increase 116% by 2025.