McAfee-Secured Website

Exam Bundle

Exam Code: CAS-005

Exam Name CompTIA SecurityX

Certification Provider: CompTIA

Corresponding Certification: CASP

CompTIA CAS-005 Bundle $44.99

100% CompTIA CAS-005 Guaranteed Success With Testking Training Materials

Get Real CAS-005 Exam Questions & Expert Verified Answers as Experienced on the Real Exam!

  • Questions & Answers

    CAS-005 Questions & Answers

    482 Questions & Answers

    The ultimate exam preparation tool, Questions & Answers mirrors the actual CAS-005 exam while allowing you to hone and measure every aspect of your test-taking ability.

  • CAS-005 Video Course

    CAS-005 Video Course

    180 Video Lectures

    CAS-005 Video Course is developed by CompTIA Professionals to help you pass the CAS-005 exam.

    Description

    This course will improve your knowledge and skills required to pass CompTIA SecurityX exam.
  • Study Guide

    CAS-005 Study Guide

    903 PDF Pages

    Developed by industry experts, this 903-page guide spells out in painstaking detail all of the information you need to ace CAS-005 exam.

CompTIA SecurityX (CAS-005) Practice Tests

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Total Cost: $194.97
Bundle Price: $149.98

Purchase Individually

  • Questions & Answers

    Questions & Answers

    482 Questions

    $124.99
  • CAS-005 Video Course

    Video Course

    180 Video Lectures

    $39.99
  • Study Guide

    Study Guide

    903 PDF Pages

    $29.99

CompTIA CAS-005 Practice Test 12

Question 110

Which test or documented result best shows that the underlying issue was properly addressed? An internal coding assistant sends proprietary source files to a public model endpoint despite a restricted-data policy.

  1. Test uploads of restricted repositories are blocked and logged without transferring code externally.
  2. Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.
  3. A release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets.
  4. Access is blocked pending an authorized cross-border decision and supporting jurisdictional documentation.

Explanation

Correct Answer

A:

This is direct test evidence of the control objective in the case: An internal coding assistant sends proprietary source files to a public model endpoint despite a restricted-data policy. It demonstrates that the proposed response, enforce approved model endpoints, source-code classification checks, and egress control on AI integrations., produces the measurable result test uploads of restricted repositories are blocked and logged without transferring code externally. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

B:

For the issue “An internal coding assistant sends proprietary source files to a public model endpoint despite a restricted-data policy”, measuring “Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.” would validate “A model-training pipeline includes customer identifiers that are not necessary for the intended forecasting task” instead of the required control effect: test uploads of restricted repositories are blocked and logged without transferring code externally.

C:

This measurement is tied to “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced” and the control “Require supplier provenance attestations, artifact integrity, and test-data chain of custody for model release”. It cannot confirm that “Enforce approved model endpoints, source-code classification checks, and egress control on AI integrations” has achieved “Test uploads of restricted repositories are blocked and logged without transferring code externally” for the issue in this question.

D:

Observation “Access is blocked pending an authorized cross-border decision and supporting jurisdictional documentation.” would be useful for “An international supplier proposes to move restricted research telemetry to a foreign support location without export review”, but it does not prove that the response to “An internal coding assistant sends proprietary source files to a public model endpoint despite a restricted-data policy” achieved the relevant outcome, test uploads of restricted repositories are blocked and logged without transferring code externally.

Question 111

In a control effectiveness review, the issue is: A document summarizer's plugin can read any employee mailbox, even though users request summaries of their own threads. Which intervention is most defensible for the situation?

  1. Constrain agents to allowlisted actions and require approval for destructive or privileged execution.
  2. Scope plugin tokens to requester-authorized resources and enforce per-action user authorization.
  3. Establish a common risk-based change standard while recording plant-specific safety authorization requirements.
  4. Enforce approved model endpoints, source-code classification checks, and egress control on AI integrations.

Explanation

Correct Answer

B:

Scope plugin tokens to requester-authorized resources and enforce per-action user authorization. The acceptance check is that a test request for another employee's messages is denied at the mailbox permission boundary. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

“Constrain agents to allowlisted actions and require approval for destructive or privileged execution” addresses “An enterprise faces this security constraint: A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization”, with “An injected ticket command is rejected by tool policy before reaching a shell” as its acceptance test. Here the specific problem is “A document summarizer's plugin can read any employee mailbox, even though users request summaries of their own threads”: it requires “Scope plugin tokens to requester-authorized resources and enforce per-action user authorization” and evidence that “A test request for another employee's messages is denied at the mailbox permission boundary”.

C:

For “A manufacturer uses different firewall-change approval rules at headquarters and its newly acquired plant”, “Establish a common risk-based change standard while recording plant-specific safety authorization requirements” has a clear validation signal: “Both locations show approved changes with consistent risk reviews and documented safety signoffs where required”. But “A document summarizer's plugin can read any employee mailbox, even though users request summaries of their own threads” calls for “Scope plugin tokens to requester-authorized resources and enforce per-action user authorization”; the other intervention cannot demonstrate “A test request for another employee's messages is denied at the mailbox permission boundary”.

D:

“Test uploads of restricted repositories are blocked and logged without transferring code externally” would be meaningful evidence for “Enforce approved model endpoints, source-code classification checks, and egress control on AI integrations” in the case “A cross-functional security group is considering a decision: An internal coding assistant sends proprietary source files to a public model endpoint despite a restricted-data policy”. It does not replace “Scope plugin tokens to requester-authorized resources and enforce per-action user authorization” to correct “A document summarizer's plugin can read any employee mailbox, even though users request summaries of their own threads” and verify “A test request for another employee's messages is denied at the mailbox permission boundary”.

Question 112

A follow-up assurance review considers this problem: A document summarizer's plugin can read any employee mailbox, even though users request summaries of their own threads. Which outcome best shows the intended improvement?

  1. Random sampled interactions include the required AI-use notice and audit records.
  2. Contract administrators correctly reject a simulated payment-change request lacking dual approval.
  3. A test request for another employee's messages is denied at the mailbox permission boundary.
  4. Test uploads of restricted repositories are blocked and logged without transferring code externally.

Explanation

Correct Answer

C:

This is direct test evidence of the control objective in the case: A document summarizer's plugin can read any employee mailbox, even though users request summaries of their own threads. It demonstrates that the proposed response, scope plugin tokens to requester-authorized resources and enforce per-action user authorization., produces the measurable result a test request for another employee's messages is denied at the mailbox permission boundary. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In “A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy”, “Random sampled interactions include the required AI-use notice and audit records” is the acceptance signal for “Define disclosure, retention, and oversight requirements for AI-mediated customer communications”. The required proof for the selected response “Scope plugin tokens to requester-authorized resources and enforce per-action user authorization” is “A test request for another employee's messages is denied at the mailbox permission boundary”.

B:

“Contract administrators correctly reject a simulated payment-change request lacking dual approval” would verify “Develop a risk-based training matrix covering supplier-payment fraud, escalation rules, and role-specific authorization” for the situation “A multinational's annual awareness program ignores contract administrators with access to supplier payment details”. Here the required evidence is “A test request for another employee's messages is denied at the mailbox permission boundary” after “Scope plugin tokens to requester-authorized resources and enforce per-action user authorization”; the alternative observation does not test that control.

D:

The observed “Test uploads of restricted repositories are blocked and logged without transferring code externally” validates the separate control “Enforce approved model endpoints, source-code classification checks, and egress control on AI integrations” under “A cross-functional security group is considering a decision: An internal coding assistant sends proprietary source files to a public model endpoint despite a restricted-data policy”. The present decision calls for the different result “A test request for another employee's messages is denied at the mailbox permission boundary” from “Scope plugin tokens to requester-authorized resources and enforce per-action user authorization”.

Question 113

A security architect is asked to resolve this problem: A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases. Select the MOST directly relevant action.

  1. Add adversarial evaluation for rate exhaustion, sensitive output, prompt injection, and model extraction.
  2. Enforce approved model endpoints, source-code classification checks, and egress control on AI integrations.
  3. Threat-model extension authority, data exfiltration paths, and indirect prompt injection across trust boundaries.
  4. Implement out-of-band verification and dual approval for sensitive transfer instructions regardless of voice similarity.

Explanation

Correct Answer

A:

Add adversarial evaluation for rate exhaustion, sensitive output, prompt injection, and model extraction. The acceptance check is that a test suite records controlled failure rates, safeguards, and remediation owners across attack categories. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

B:

Although “Enforce approved model endpoints, source-code classification checks, and egress control on AI integrations.” is a reasonable response to “An internal coding assistant sends proprietary source files to a public model endpoint despite a restricted-data policy”, this case instead concerns “A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases”, for which success requires that a test suite records controlled failure rates, safeguards, and remediation owners across attack categories.

C:

The option “Threat-model extension authority, data exfiltration paths, and indirect prompt injection across trust boundaries.” assumes an issue like “An AI analytics product processes privileged customer documents using a new extension with unscreened API permissions”, not the specific observed weakness “A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases”; it leaves the evidence goal unmet: a test suite records controlled failure rates, safeguards, and remediation owners across attack categories.

D:

The proposal “Implement out-of-band verification and dual approval for sensitive transfer instructions regardless of voice similarity” manages “A marketing team uses synthetic executive audio to authorize an urgent bank transfer through a voice call”, where successful operation means “A simulated convincing synthetic call cannot authorize transfer without independent second-channel confirmation”. The present issue instead needs “Add adversarial evaluation for rate exhaustion, sensitive output, prompt injection, and model extraction” to establish “A test suite records controlled failure rates, safeguards, and remediation owners across attack categories”.

Question 114

After the team responds to the following issue, which finding best demonstrates the desired security outcome? A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases.

  1. Threat records show assets, attacker paths, control owners, and corresponding security test cases.
  2. Random sampled interactions include the required AI-use notice and audit records.
  3. A test suite records controlled failure rates, safeguards, and remediation owners across attack categories.
  4. Malicious document instructions fail to trigger privileged tool calls in a controlled evaluation.

Explanation

Correct Answer

C:

This is direct test evidence of the control objective in the case: A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases. It demonstrates that the proposed response, add adversarial evaluation for rate exhaustion, sensitive output, prompt injection, and model extraction., produces the measurable result a test suite records controlled failure rates, safeguards, and remediation owners across attack categories. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

A:

In the problem “A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases”, evidence “Threat records show assets, attacker paths, control owners, and corresponding security test cases.” is not decisive because it tests “A team lists spoofing and repudiation threats in a STRIDE spreadsheet but lacks specific affected components”; the requested validation is a test suite records controlled failure rates, safeguards, and remediation owners across attack categories.

B:

For the issue “A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases”, measuring “Random sampled interactions include the required AI-use notice and audit records.” would validate “A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy” instead of the required control effect: a test suite records controlled failure rates, safeguards, and remediation owners across attack categories.

D:

Observation “Malicious document instructions fail to trigger privileged tool calls in a controlled evaluation.” would be useful for “A retrieval-augmented assistant treats instructions hidden in supplier PDF content as higher-priority system commands”, but it does not prove that the response to “A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases” achieved the relevant outcome, a test suite records controlled failure rates, safeguards, and remediation owners across attack categories.

Question 115

A senior engineer examines the following finding: A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced. Which action most directly addresses the requirement?

  1. Add adversarial evaluation for rate exhaustion, sensitive output, prompt injection, and model extraction.
  2. Constrain agents to allowlisted actions and require approval for destructive or privileged execution.
  3. Model the supply-chain compromise path and design an independent package signing and verification boundary.
  4. Require supplier provenance attestations, artifact integrity, and test-data chain of custody for model release.

Explanation

Correct Answer

D:

Require supplier provenance attestations, artifact integrity, and test-data chain of custody for model release. The acceptance check is that a release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

The option “Add adversarial evaluation for rate exhaustion, sensitive output, prompt injection, and model extraction.” assumes an issue like “A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases”, not the specific observed weakness “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced”; it leaves the evidence goal unmet: a release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets.

B:

This choice would be evaluated by “An injected ticket command is rejected by tool policy before reaching a shell” when dealing with “An enterprise faces this security constraint: A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization”. It does not deliver the distinct control “Require supplier provenance attestations, artifact integrity, and test-data chain of custody for model release” needed for “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced”.

C:

“A simulated modified firmware package is rejected before installation” would be meaningful evidence for “Model the supply-chain compromise path and design an independent package signing and verification boundary” in the case “A supplier's firmware update path allows unsigned packages from a third-party hosting mirror”. It does not replace “Require supplier provenance attestations, artifact integrity, and test-data chain of custody for model release” to correct “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced” and verify “A release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets”.

Question 116

Which test or documented result best shows that the underlying issue was properly addressed? A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced.

  1. A release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets.
  2. Malicious document instructions fail to trigger privileged tool calls in a controlled evaluation.
  3. The scoped model removes unsupported attack paths and prioritizes credible techniques with evidence.
  4. Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.

Explanation

Correct Answer

A:

This is direct test evidence of the control objective in the case: A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced. It demonstrates that the proposed response, require supplier provenance attestations, artifact integrity, and test-data chain of custody for model release., produces the measurable result a release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

B:

Observation “Malicious document instructions fail to trigger privileged tool calls in a controlled evaluation.” would be useful for “A retrieval-augmented assistant treats instructions hidden in supplier PDF content as higher-priority system commands”, but it does not prove that the response to “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced” achieved the relevant outcome, a release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets.

C:

In the problem “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced”, evidence “The scoped model removes unsupported attack paths and prioritizes credible techniques with evidence.” is not decisive because it tests “A project imports a threat library but assigns high risk to every MITRE ATT&CK technique regardless of local applicability”; the requested validation is a release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets.

D:

For the issue “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced”, measuring “Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.” would validate “A model-training pipeline includes customer identifiers that are not necessary for the intended forecasting task” instead of the required control effect: a release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets.

Question 117

An enterprise faces this security constraint: A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization. Which proposal best addresses the stated risk?

  1. Require explainability, human review, and source-traceable supporting indicators for consequential decisions.
  2. Evaluate data classification, cross-border transfer rules, contractual restrictions, and required approval before moving access.
  3. Restrict tool permissions to approved refund ranges and require independent authorization for high-value operations.
  4. Constrain agents to allowlisted actions and require approval for destructive or privileged execution.

Explanation

Correct Answer

D:

Constrain agents to allowlisted actions and require approval for destructive or privileged execution. The acceptance check is that an injected ticket command is rejected by tool policy before reaching a shell. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

The proposal “Require explainability, human review, and source-traceable supporting indicators for consequential decisions” manages “An executive uses a generative AI risk score to deny supplier access without understanding confidence or evidence”, where successful operation means “A sample decision includes verified underlying evidence, reviewer signoff, and an appeals process”. The present issue instead needs “Constrain agents to allowlisted actions and require approval for destructive or privileged execution” to establish “An injected ticket command is rejected by tool policy before reaching a shell”.

B:

Although “Evaluate data classification, cross-border transfer rules, contractual restrictions, and required approval before moving access.” is a reasonable response to “An international supplier proposes to move restricted research telemetry to a foreign support location without export review”, this case instead concerns “A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization”, for which success requires that an injected ticket command is rejected by tool policy before reaching a shell.

C:

The option “Restrict tool permissions to approved refund ranges and require independent authorization for high-value operations.” assumes an issue like “A support chatbot can submit payment refunds using the employee's broad API token without approval thresholds”, not the specific observed weakness “A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization”; it leaves the evidence goal unmet: an injected ticket command is rejected by tool policy before reaching a shell.

Question 118

A follow-up assurance review considers this problem: A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization. Which outcome best shows the intended improvement?

  1. An injected ticket command is rejected by tool policy before reaching a shell.
  2. A validated security diagram identifies command initiators and controls preventing unauthorized office-to-plant operations.
  3. A release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets.
  4. Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.

Explanation

Correct Answer

A:

This is direct test evidence of the control objective in the case: A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization. It demonstrates that the proposed response, constrain agents to allowlisted actions and require approval for destructive or privileged execution., produces the measurable result an injected ticket command is rejected by tool policy before reaching a shell. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

B:

For the issue “A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization”, measuring “A validated security diagram identifies command initiators and controls preventing unauthorized office-to-plant operations.” would validate “An IoT control architecture lacks a view of how operational commands cross between building automation and office IT” instead of the required control effect: an injected ticket command is rejected by tool policy before reaching a shell.

C:

“A release candidate has verifiable model hashes, supplier evidence, and traceable evaluation datasets” would verify “Require supplier provenance attestations, artifact integrity, and test-data chain of custody for model release” for the situation “A healthcare AI model's external training supplier cannot identify where checkpoint weights and evaluation data were sourced”. Here the required evidence is “An injected ticket command is rejected by tool policy before reaching a shell” after “Constrain agents to allowlisted actions and require approval for destructive or privileged execution”; the alternative observation does not test that control.

D:

Observation “Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.” would be useful for “A model-training pipeline includes customer identifiers that are not necessary for the intended forecasting task”, but it does not prove that the response to “A virtual assistant executes shell commands derived from untrusted ticket descriptions without human authorization” achieved the relevant outcome, an injected ticket command is rejected by tool policy before reaching a shell.

Question 119

An enterprise security review reports: A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy. Which response best addresses the underlying security issue?

  1. Apply data minimization, documented processing purpose, and privacy review before training.
  2. Separate retrieved documents from trusted instructions and validate tool requests against independent policy.
  3. Define retention and secure disposal standards across production, analytics, and backups with legal hold exceptions.
  4. Define disclosure, retention, and oversight requirements for AI-mediated customer communications.

Explanation

Correct Answer

D:

Define disclosure, retention, and oversight requirements for AI-mediated customer communications. The acceptance check is that random sampled interactions include the required AI-use notice and audit records. That outcome checks whether the requirement is met, rather than treating completion of a security task as proof of effectiveness.

Incorrect Answers

A:

The option “Apply data minimization, documented processing purpose, and privacy review before training.” assumes an issue like “A model-training pipeline includes customer identifiers that are not necessary for the intended forecasting task”, not the specific observed weakness “A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy”; it leaves the evidence goal unmet: random sampled interactions include the required AI-use notice and audit records.

B:

“Separate retrieved documents from trusted instructions and validate tool requests against independent policy” addresses “A retrieval-augmented assistant treats instructions hidden in supplier PDF content as higher-priority system commands”, with “Malicious document instructions fail to trigger privileged tool calls in a controlled evaluation” as its acceptance test. Here the specific problem is “A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy”: it requires “Define disclosure, retention, and oversight requirements for AI-mediated customer communications” and evidence that “Random sampled interactions include the required AI-use notice and audit records”.

C:

The proposal “Define retention and secure disposal standards across production, analytics, and backups with legal hold exceptions” manages “A healthcare service retains customer exports indefinitely because data disposal is omitted from its development lifecycle”, where successful operation means “A sample expired export is removed under approved procedure while legally held material remains protected”. The present issue instead needs “Define disclosure, retention, and oversight requirements for AI-mediated customer communications” to establish “Random sampled interactions include the required AI-use notice and audit records”.

Question 120

After the team responds to the following issue, which finding best demonstrates the desired security outcome? A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy.

  1. Random sampled interactions include the required AI-use notice and audit records.
  2. A test suite records controlled failure rates, safeguards, and remediation owners across attack categories.
  3. A sampled control supports multiple frameworks without duplicate untraceable ownership or missing required proof.
  4. Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.

Explanation

Correct Answer

A:

This is direct test evidence of the control objective in the case: A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy. It demonstrates that the proposed response, define disclosure, retention, and oversight requirements for AI-mediated customer communications., produces the measurable result random sampled interactions include the required AI-use notice and audit records. The observation tests a relevant security outcome rather than an unrelated control or reporting activity.

Incorrect Answers

B:

Observation “A test suite records controlled failure rates, safeguards, and remediation owners across attack categories.” would be useful for “A team tests an AI model using only ordinary prompts, omitting model denial-of-service and extraction abuse cases”, but it does not prove that the response to “A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy” achieved the relevant outcome, random sampled interactions include the required AI-use notice and audit records.

C:

In the problem “A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy”, evidence “A sampled control supports multiple frameworks without duplicate untraceable ownership or missing required proof.” is not decisive because it tests “An acquisition brings a second audit control library and duplicate safeguards with inconsistent descriptions”; the requested validation is random sampled interactions include the required AI-use notice and audit records.

D:

For the issue “A regulator asks whether customer support calls were AI-generated, but the organization has no disclosure policy”, measuring “Training inputs exclude unnecessary identifiers and a privacy audit traces approved data sources.” would validate “A model-training pipeline includes customer identifiers that are not necessary for the intended forecasting task” instead of the required control effect: random sampled interactions include the required AI-use notice and audit records.