Generative AI widens the attack surface, because the AI itself can be attacked, for example by prompt injection or data poisoning. Prompt injection changes the input so a system behaves in ways nobody intended; in indirect prompt injection, the instructions hide in data the system is likely to read. Data poisoning tampers with training data to change a model's outputs. A secure system blocks unauthorized access, a resilient one withstands surprises or fails safely, and a safe one does not endanger people, with safety starting at design. AI red-teaming is a structured test to find flaws, and everyday users can red-team too.