Rethinking Risk in the Age of AI

**Executive summary:** This HackerOne briefing outlines how generative AI expands attack surfaces and enumerates AI-specific threats (prompt injection, jailbreaking, dependency attacks, data poisoning, denial-of-service, misuse), highlights lessons from Anthropic and Snap using crowdsourced red teaming, and recommends integrating security into AI development lifecycles, combining human expertise with automation, and leveraging external researchers to improve resilience.