An organization is developing and implementing AI features into a customer service application.

Basic Concept: Before deploying AI applications that handle customer data in trials, protecting sensitive information through data masking and sanitization is essential. CompTIA SecAI+ Study Guide emphasizes pre-deployment data security controls as a critical step in the AI development lifecycle.

Why A is Correct: Data masking replaces sensitive real customer data with realistic but fictitious equivalents, while sanitization removes harmful or unwanted data elements. Before customer trials, these techniques prevent exposure of real PII or sensitive information, ensure the trial environment cannot leak production data, and protect the organization from privacy regulation violations. This is the most immediately actionable pre-trial security control.

Why B is Wrong: External compliance audits are formal processes typically conducted post-deployment or at planned intervals to verify regulatory compliance. They are not pre-trial security implementations and cannot prevent data exposure in a trial environment.

Why C is Wrong: Approved AI vendor lists are governance artifacts that manage vendor selection risk at the procurement stage. They do not directly protect customer data within an application being prepared for trials.

Why D is Wrong: Third-party risk management addresses risks from external vendors and partners at a strategic level. While important for overall governance, it does not constitute a direct data security control for a pre-trial release.