CCAR-P Governance, Safety, and Risk Management Practice Question
An organization is conducting a risk assessment for an AI application. They are concerned about 'model drift' over time. Which governance action is most appropriate to manage this risk?
⚠ Common exam trap
Candidates often mistake model drift for a security breach or a training data issue. They focus on retraining the model immediately rather than implementing a monitoring process to detect the degradation first.
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
✓
Conduct periodic evaluation against a golden dataset.
Model drift refers to the degradation of model performance as the environment or data distribution changes. Periodic re-evaluation against a baseline 'golden dataset' is a standard governance practice to ensure the model remains reliable. This proactive monitoring allows teams to detect performance drops, adjust system prompts, or re-train/update the model, ensuring that the AI remains safe and effective for its original intended use case.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Increase the system prompt complexity.
Why it's wrong here
Increasing prompt complexity generally makes the model harder to evaluate and can introduce new, unpredictable behaviors. It does not address the underlying issue of performance drift caused by changing input data distributions. Governance requires measurable evaluation, not just adjusting instructions to try and 'fix' behavior after the fact.
- ✓
Conduct periodic evaluation against a golden dataset.
Why this is correct
A golden dataset provides a consistent benchmark to measure the model's performance over time. Comparing current outputs against this baseline allows organizations to quantify drift, identify specific areas of failure, and maintain quality control. This is the most effective way to ensure long-term stability in AI production environments.
- ✗
Switch to a different model provider immediately.
Why it's wrong here
Switching providers without a root cause analysis is reactive and inefficient. Drift is a common challenge in all LLM deployments. The correct governance approach is to measure and manage the existing system performance rather than simply assuming a different provider will solve issues without proper validation and testing protocols.
- ✗
Automate user feedback loops for real-time retraining.
Why it's wrong here
Automating retraining based on real-time feedback introduces massive risks, as users can inadvertently (or intentionally) inject bias or poor-quality data into the model. Controlled, offline evaluation cycles are the industry standard for safe AI governance; real-time retraining lacks the necessary human oversight and quality assurance required for enterprise safety.
About these practice questions
One of 262 original CCAR-P practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. 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-P 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-P exam.