Aegentra Labs

Free, no account required

ISO/IEC 42001 Foundation practice questions

Five stand-alone questions, as in the Foundation exam: AI stakeholder roles, impact versus risk assessment, measurable AI objectives, assessment after significant change and supplier responsibilities. Reveal each answer to see the reasoning and why the other options fail.

These practice questions are independently authored by Harry Sidhu and Aegentra and are not official PECB examination materials. They do not reproduce confidential or recalled PECB examination content. Read the methodology.

Questions
5
Options
3 each
Answers
Explained
Start question 1

What this exact sample covers

AIMS concepts applied

  • AI roles and impact assessment
  • AI objectives and change
  • Annex A responsibilities
  1. Question 01Domain 1 · application

    A regional bank buys a speech-analytics service from a software vendor that built and trained the model. The bank uses the service to score recorded complaint calls for urgency, and callers are told their calls may be analysed. In the AI stakeholder roles that ISO/IEC 42001:2023 draws from ISO/IEC 22989, which role do the callers hold?

    1. AAI users, because their recorded speech is the input that the speech-analytics service processes to produce each urgency score.
    2. BAI subjects, because they are affected by the system's outputs even though they neither operate the service nor contract for it.
    3. CAI customers, because the urgency scores are produced to improve the complaint handling that the bank delivers to these callers.
    the answer and reasoning

    Correct answerB. AI subjects, because they are affected by the system's outputs even though they neither operate the service nor contract for it.

    ISO/IEC 22989 describes an AI subject as an entity affected by an AI system, such as a data subject whose data is processed. The callers' speech is analysed and their complaints are prioritised by the output, but they do not use or buy the service. The bank is the AI customer and its staff are AI users; the vendor is the AI provider and producer.

    Why the other options fail

    • A. AI users interact with or operate the system; supplying data that the system processes makes the callers data subjects, a kind of AI subject, rather than users.
    • C. The AI customer is the entity that obtains and uses the AI product or service, here the bank; benefiting from better complaint handling does not make callers customers.

    Published exam-scope sources: ISO/IEC 42001:2023 artificial intelligence management systems · ISO/IEC 42001 Foundation training course page

  2. Question 02Domain 1 · analysis

    An energy retailer plans an AI system that flags households for suspected meter tampering. Its AI risk register rates consequences for the retailer, such as revenue loss and regulatory penalties. A reviewer asks how flagging could affect elderly customers living alone who are visited by investigators. Which process should the retailer use to examine this, with its results then considered in the risk assessment?

    1. AAI risk treatment under clause 6.1.3, because selecting controls for the visits will resolve the concern raised by the reviewer about elderly customers.
    2. BManagement review under clause 9.3, because top management must decide whether the effects on vulnerable customers are acceptable to the retailer.
    3. CAI system impact assessment under clause 6.1.4, because it determines potential consequences of the system for individuals, groups and societies.
    the answer and reasoning

    Correct answerC. AI system impact assessment under clause 6.1.4, because it determines potential consequences of the system for individuals, groups and societies.

    Clause 6.1.4 requires a process for assessing the potential consequences that the development, provision or use of AI systems can have for individuals, groups of individuals and societies, and its results are considered in the AI risk assessment under clause 6.1.2. The reviewer's concern is exactly such a consequence for a group of people, so it is examined there before controls are chosen.

    Why the other options fail

    • A. Risk treatment selects controls for risks that have already been assessed; it does not examine the consequences for affected people that the reviewer has raised.
    • B. Management review evaluates the AIMS at planned intervals; it is not the process for determining an AI system's consequences for particular groups of people.

    Published exam-scope sources: ISO/IEC 42001:2023 artificial intelligence management systems · ISO/IEC 42001 Foundation training course page

  3. Question 03Domain 2 · application

    Top management of a rental-housing platform approves the AI objective ‘Our application-ranking model will treat applicants fairly’. The platform already measures monthly approval-rate ratios across applicant groups. Does this objective meet clause 6.2 of ISO/IEC 42001:2023?

    1. ANo, because AI objectives are to be measurable where practicable, and a suitable measure for this objective already exists.
    2. BYes, because AI objectives are only required to be consistent with the AI policy, and fairness is one of the policy's principles.
    3. CYes, because clause 6.2 leaves measurement to monitoring under clause 9.1 rather than requiring it in the objective itself.
    the answer and reasoning

    Correct answerA. No, because AI objectives are to be measurable where practicable, and a suitable measure for this objective already exists.

    Clause 6.2 requires AI objectives to be consistent with the AI policy, measurable if practicable, monitored, communicated, updated as appropriate and documented. Here measurement is clearly practicable because approval-rate ratios are already produced every month, so a purely qualitative statement of fairness does not meet the clause. Consistency with the policy is necessary but not sufficient.

    Why the other options fail

    • B. Consistency with the AI policy is only one of several criteria in clause 6.2; objectives must also be measurable where practicable, and here they can be.
    • C. Clause 6.2 itself requires objectives to be measurable if practicable; clause 9.1 monitoring depends on measurable objectives rather than replacing that requirement.

    Published exam-scope sources: ISO/IEC 42001:2023 artificial intelligence management systems · ISO/IEC 42001 Foundation training course page

  4. Question 04Domain 2 · evaluation

    A food-delivery platform assesses AI risks every twelve months, and the next assessment is due in nine months. It now plans to let its courier-rating model deactivate couriers automatically, without the current human review, from next month. What does ISO/IEC 42001:2023 require before this change?

    1. AProceed and include the change in the next scheduled assessment, because the organization sets the planned intervals for its AI risk assessments.
    2. BProceed once the Statement of Applicability is updated, because removing human review changes a control rather than the AI risks themselves.
    3. CAssess the AI risks and the AI system impacts now, because significant proposed changes trigger these assessments as well as planned intervals.
    the answer and reasoning

    Correct answerC. Assess the AI risks and the AI system impacts now, because significant proposed changes trigger these assessments as well as planned intervals.

    Clauses 8.2 and 8.4 require AI risk assessments and AI system impact assessments at planned intervals or when significant changes are proposed. Removing human review from decisions that end couriers' work is a significant change to how the system affects people, so both assessments are due before it takes effect. Waiting nine months or only editing the SoA ignores the change trigger.

    Why the other options fail

    • A. Planned intervals are only one trigger; clauses 8.2 and 8.4 also require assessments when significant changes are proposed, as this one is.
    • B. Removing a human oversight control changes the level of AI risk and the impact on couriers, so the SoA cannot be updated without first reassessing.

    Published exam-scope sources: ISO/IEC 42001:2023 artificial intelligence management systems · ISO/IEC 42001 Foundation training course page

  5. Question 05Domain 2 · analysis

    A logistics company deploys a supplier's AI document-extraction service in its customs process. The supplier releases a new model version each month, and the contract is silent on who checks extraction accuracy after each release. Which Annex A control of ISO/IEC 42001:2023 most directly addresses this gap?

    1. AControl A.4.4, because the supplier's model is a tooling resource that the company has to document as part of its AI system resources.
    2. BControl A.10.2, because responsibilities across the AI system life cycle are to be allocated between the company and its suppliers.
    3. CControl A.6.2.7, because the supplier has to provide technical documentation for each model version that it releases to customers.
    the answer and reasoning

    Correct answerB. Control A.10.2, because responsibilities across the AI system life cycle are to be allocated between the company and its suppliers.

    Control A.10.2 requires the organization to ensure that responsibilities within the AI system life cycle are allocated between itself, its partners, suppliers, customers and third parties. Here nobody has been assigned to check accuracy after monthly releases, which is a gap in allocated responsibility. Documenting resources or receiving technical documentation would not settle who performs the check.

    Why the other options fail

    • A. Control A.4.4 concerns documenting tooling resources used for AI systems; recording the supplier's model as a resource does not assign anyone to verify accuracy after releases.
    • C. Control A.6.2.7 concerns technical documentation of the AI system; receiving documentation for each version still leaves open who checks the accuracy of the output.

    Published exam-scope sources: ISO/IEC 42001:2023 artificial intelligence management systems · ISO/IEC 42001 Foundation training course page

Continue preparing

Ready for the full Foundation bank?

Casual mode shows feedback as you answer. Full access adds the complete weighted bank, exam mode, saved results and domain-level analysis for this exact qualification. Start with a 3-day free trial; practice sessions are saved to your account.

Preparing for a different exam?

Every live exam has its own five free questions, written for that exam's domains.