Generative AI systems are vulnerable to attacks via natural language, with 35% of real-world AI security incidents caused by simple prompts, highlighting the need for cybersecurity practitioners to adapt their skills to this new threat landscape
