Question

A healthcare provider wants to build a patient-facing chatbot that answers questions about appointment scheduling and general wellness. Legal has mandated that the application must never expose or store any protected health information (PHI) and must refuse to provide specific medical diagnoses. As the GenAI engineer, which design approach best translates these business requirements into the application architecture?

Accepted Answer

Add an input filtering layer that detects and redacts PHI before it reaches the LLM, plus an output guardrail that blocks diagnostic responses and constrains the model to scheduling/wellness scope via system prompt.

Answer

Fine-tune the base model on the provider's historical patient chat logs so it learns to answer scheduling and wellness questions accurately.

Answer

Increase the model's context window so it can retain the entire conversation and rely on the LLM to self-censor any PHI or diagnostic content.

Answer

Log all conversations to a vector database for later retrieval and audit, then instruct the model in the prompt to avoid diagnoses.