Your question is Mitigating AI-Specific Threats. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
What strategies do you use to address AI-specific risks such as prompt injection, data exposure, and output manipulation?
Explain how you would threat-model an AI-powered Adobe application and design practical preventive, detective, and responsive controls. Address untrusted prompts, retrieval data, tool access, model outputs, identity, logging, and incident response.