Your question is LLM Enterprise Security Concerns. 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 are the primary security concerns when deploying LLMs in an enterprise environment like New York Life?
Explain how you would assess and prioritize risks across the model, prompts, retrieval data, tools, infrastructure, and user-facing application. Include practical controls for prompt injection, data leakage, hallucinations, excessive agency, access control, monitoring, and incident response. Describe how you would validate the controls with offline and online evaluation, while considering latency and operating cost.