CCAR-F Context and Reliability Practice Question
A legal team uses Claude to summarize case files. They require that the summary never includes personally identifiable information (PII) such as names, addresses, or phone numbers. The case files are lengthy and contain PII in various formats. Which approach is most reliable to ensure PII is not included in the summary?
⚠ Common exam trap
The trap here is trusting the model's instruction-following or fine-tuning to reliably redact PII, when a deterministic pre-processing step is the only way to guarantee removal.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Use a named entity recognition (NER) tool to detect and redact PII from the case files before sending them to Claude.
The most reliable way to prevent PII in summaries is to remove it before Claude processes the text. A dedicated NER tool can be configured for high recall and deterministic redaction, ensuring that no PII reaches the model. System prompt instructions and fine-tuning are probabilistic and may miss instances, while manual review is impractical at scale.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Fine-tune Claude on a dataset of legal documents with PII already redacted.
Why it's wrong here
Fine-tuning can teach the model patterns of redaction, but it does not guarantee that all PII will be removed, especially unseen formats. It is also costly and time-consuming to update. For a strict compliance requirement, a deterministic tool is more reliable. Fine-tuning may improve performance but cannot replace a dedicated redaction step.
- ✗
Ask Claude to output the summary in a structured format with a separate field for PII, then manually review it.
Why it's wrong here
This approach still requires Claude to identify PII, which is error-prone, and adds a manual review step that is not scalable. It also risks PII appearing in the summary field if Claude misclassifies. Manual review can catch errors but is not reliable for large volumes. A deterministic pre-processing step is superior.
- ✓
Use a named entity recognition (NER) tool to detect and redact PII from the case files before sending them to Claude.
Why this is correct
Pre-processing with a dedicated NER tool ensures that PII is removed deterministically before the text reaches Claude. This approach does not rely on the model's compliance and can be tuned for high recall. It is the most reliable method because it addresses the problem at the source, and Claude then summarizes only the redacted text, eliminating the risk of PII leakage in the summary.
- ✗
Instruct Claude in the system prompt to redact all PII from the summary.
Why it's wrong here
While a system prompt instruction is helpful, it is not reliable on its own because Claude may miss PII in complex or ambiguous formats. LLMs can inadvertently include PII if it appears in contexts they deem relevant. For a strict requirement like PII redaction, a deterministic pre-processing step is more reliable. Instructions alone cannot guarantee 100% removal.
About these practice questions
This CCAR-F question is part of Courseiva's 271-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Anthropic exam blueprint
This CCAR-F practice question is part of Courseiva's free Anthropic certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the CCAR-F exam.