Exam CCDV-F Topic 1 Question 82 Discussion

Option B establishes the correct trust boundary. Fluent or confident language is not evidence that a generated claim is factually correct. If downstream systems treat output as authoritative data, the application must independently establish whether the output meets its correctness requirements before accepting it.
Validation can take several forms depending on the workload: compare generated facts against authoritative records, require citations or source references, constrain output to retrieved evidence, apply deterministic business rules, or use separate evaluation/classification stages. Anthropic's agent engineering guidance repeatedly emphasizes explicit evaluation criteria and validation rather than relying on apparent confidence.
A confuses sampling behavior with factual reliability. Lowering temperature does not establish factual correctness and may only make an incorrect answer more repeatable. C supplies maximum oversight but is unnecessarily expensive and removes useful automation even for low-risk, easily validated cases. D communicates uncertainty to users but does not protect downstream systems that automatically consume the response.
Therefore, B treats model output as untrusted until verified to the level required by the application. Relevant Study Guide topics: output validation, grounding, factuality, confidence calibration, source verification, trust boundaries, and downstream safety.