Deep explanation
Secure Google Gemini in a customer onboarding assistant
Junior security question: threat-model a Gemini onboarding assistant for prompt injection, PII leakage, and tool misuse, and fix it with server-side authorization, redaction, Vertex AI controls, and testing.
Quick answer
I would list what the assistant can read, what it can do, and who can influence its input, because uploaded documents and chat messages are both untrusted. The model should never decide on its own to create or change accounts; our backend checks every tool call against the applicant's session and business rules. I would also limit personal data sent to Gemini, run it through Vertex AI with IAM and proper data terms, and test with injection attacks before launch.
The interview question
Your customer onboarding assistant uses Google Gemini to read uploaded ID documents and forms, answer applicant questions, and call tools that create accounts. A security review flagged prompt injection, leakage of personal data, and unauthorized tool calls. How would you assess the risks and secure the system?
Deep explanation
Sign in to unlock the full answer
Free accounts include 5 full deep answers. Sign in to start unlocking.
- 22 more sections of deep explanation
- Real-world examples
- Common mistakes
- Interviewer expectations
- Follow-up questions