30 Questions — ITIL 5 – Foundation : Certified ITIL Foundation - Domain 6 - ITIL and Artificial Intelligence

Expand any question to reveal the correct answer and explanation.

1. 1 Which specific '6C' AI capability is responsible for translating technical log data into a format that stakeholders can understand to assess business impact?

   This capability bridges the gap between raw data and human-centric understanding.

   Clarification

   Clarification focuses on simplifying or explaining complex technical information, such as logs, for broader organizational use.

   • ✗ Communication

     While related, communication refers more specifically to the natural language interaction between the AI and human stakeholders.

   • ✗ Curation

     Curation involves filtering and organizing existing information, such as correlating alerts, rather than explaining it.

   • ✗ Cognition

     Cognition is centered on high-level decision-making, predictions, and forecasting based on data patterns.

2. 2 Under the ITIL 5 AI Governance Model, which perspective addresses the authority to approve an AI agent's ability to trigger an automated remediation script?

   Consider which perspective deals specifically with 'who decides' and the limits of autonomy.

   Decision Authority and Risk Management

   This perspective defines who has the power to approve autonomous actions and sets operational boundaries for those agents.

   • ✗ Ethical Principles and Responsible AI

     This perspective is more concerned with bias, fairness, and alignment with organizational values.

   • ✗ Regulatory Compliance and Operational Standards

     While it touches on operational standards, its primary focus is on external legal frameworks like the EU AI Act.

   • ✗ Data Governance and Performance Management

     This perspective monitors model drift and training data quality rather than the delegation of authority.

3. 3 An organization discovers that its marketing team is using a third-party GenAI tool without formal IT approval to draft campaign copy. According to ITIL 5, what is the primary structural driver of this 'Shadow AI'?

   Think about the friction between business speed and governance oversight.

   Inefficient internal approval processes

   ITIL 5 identifies that slow internal governance or approval workflows are the most common cause for employees seeking unauthorized tools.

   • ✗ Intentional avoidance of ethical standards

     Source material indicates that shadow AI is rarely malicious and is instead usually a symptom of a need for speed.

   • ✗ A lack of advanced engineering resources

     The adoption of external SaaS AI tools often happens precisely because they do not require internal engineering resources.

   • ✗ High costs of approved enterprise models

     While cost can be a factor, the primary barrier cited is the agility gap created by legacy governance structures.

4. 4 Which characteristic fundamentally distinguishes Agentic AI from Generative AI in the context of ITIL 5?

   Look for the difference between 'creating content' and 'performing actions'.

   The ability to execute multi-step tasks autonomously

   Agentic AI is defined by its capacity to act across a sequence of tasks to achieve a goal without per-action human approval.

   • ✗ The capacity to produce natural language responses

     Both Generative and Agentic AI can utilize natural language, so this is not a distinguishing factor.

   • ✗ The use of large language models for reasoning

     Both types often rely on the same underlying model architectures for their processing.

   • ✗ The focus on data privacy and encryption

     Privacy and encryption are standard requirements for all enterprise AI and do not define the specific type of intelligence.

5. 5 In the 6C model, which capability is most likely involved in predicting which server is most likely to fail based on current resource utilization trends?

   This capability involves using data to understand what might happen next.

   Cognition

   Cognition encompasses predictive analytics, forecasting, and suggesting root causes based on data.

   • ✗ Curation

     Curation is about organizing and deduplicating current data alerts rather than predicting future state changes.

   • ✗ Coordination

     Coordination involves the orchestration of workflows and routing, not the analytical prediction of failures.

   • ✗ Clarification

     Clarification focuses on explaining current data rather than generating forecasts or predictive insights.

6. 6 The ITIL 5 AI Governance perspective 'Data Governance and Performance Management' primarily aims to mitigate which risk?

   This involves the ongoing monitoring of the model's 'health' and accuracy.

   Model performance drift

   This perspective monitors how models evolve over time and ensures that the data they ingest maintains quality and accuracy.

   • ✗ Algorithmic bias against users

     While data-related, bias mitigation is specifically categorized under the Ethical Principles perspective.

   • ✗ Unauthorized budget expenditure

     Financial oversight is a broader management function and not the primary focus of this specific governance perspective.

   • ✗ Lack of human-in-the-loop triggers

     Approval thresholds and human escalation fall under the Decision Authority and Risk Management perspective.

7. 7 Which of the following describes the 'Stewardship' level of the ITIL 5 AI Governance maturity model?

   It is the level where governance is seen as a supportive responsibility rather than just a set of guardrails.

   Active support for people to navigate compliance and implement measures

   Stewardship represents the highest maturity where governance actively assists stakeholders rather than just enforcing rules.

   • ✗ Formalized policies and designated executive owners for all AI

     This describes lower levels of maturity, such as the Structured Framework or Operating Model stages.

   • ✗ Ad hoc use of AI tools without centralized oversight

     This is Level 1 maturity, where AI use is uncoordinated and risky.

   • ✗ Successful pilot programs with limited documentation

     This corresponds to the Controlled Experiments level of maturity.

8. 8 Under the 6C model, 'Curation' is specifically used in incident management for which purpose?

   This capability focuses on sorting through data to find what is relevant.

   Correlating alerts and reducing operational noise

   Curation is the act of organizing raw data, such as deduplicating monitoring alerts to find the primary incident.

   • ✗ Triggering a rollback script automatically

     Triggering actions and orchestrating workflows is the domain of the Coordination capability.

   • ✗ Summarizing an incident for a post-mortem report

     Generating new text or summaries falls under the Creation capability.

   • ✗ Paging the on-call engineer based on expertise

     Routing and assigning tasks is part of the Coordination capability.

9. 9 Which AI governance perspective would be most concerned with ensuring an AI recruitment tool does not favor one demographic over another?

   Look for the perspective that focuses on fairness and social values.

   Ethical Principles and Responsible AI

   This perspective manages fairness and bias testing to ensure AI alignment with organizational ethics.

   • ✗ Decision Authority and Risk Management

     This perspective is more about who approves the system rather than the inherent fairness of its logic.

   • ✗ Data Governance and Performance Management

     This deals with the quality and drift of data, while the social impact of the logic is an ethical concern.

   • ✗ Regulatory Compliance

     While regulations may mandate fairness, the internal focus on organizational values belongs to the Ethical perspective.

10. 10 How does ITIL 5 define 'Narrow AI' in comparison to other types?

    This type of AI is focused on a 'single-purpose' function.

    AI designed to perform a specific, limited set of tasks

    Narrow AI is task-specific, whereas Generative or Agentic AI often handle more creative or autonomous multi-step operations.

    • ✗ AI that has achieved human-level general intelligence

      Human-level intelligence (AGI) is currently theoretical and not how ITIL 5 classifies current systems.

    • ✗ AI that primarily acts without any human oversight

      This describes high-autonomy Agentic AI, not necessarily the scope of the AI's intelligence.

    • ✗ AI that is only used for data visualization

      Data visualization is just one possible use case; the term refers to the specificity of the function.

11. 11 According to the ITIL 5 AI Governance Model, what is the role of 'Human-in-the-loop' (HITL) oversight?

    It serves as a checkpoint for critical decisions made by an algorithm.

    A mechanism for review thresholds in high-risk systems

    HITL is a sign of mature governance where humans must review or approve certain consequential AI decisions.

    • ✗ A fallback for when AI models fail to generate an output

      HITL is a proactive governance requirement for risk management, not just a technical redundancy.

    • ✗ A method to train the AI by manual data entry

      While humans do train AI, 'oversight' specifically refers to the control and monitoring of live systems.

    • ✗ A temporary measure only used during the pilot stage

      HITL is an ongoing operational requirement for systems classified as high-risk.

12. 12 Which of the following is a step in the ITIL AI Governance Improvement Model?

    The first step involves evaluating the current state and identifying breaking points.

    Assess and Stress-Test

    The model's four steps include Assess, Design, Implement, and Maintain.

    • ✗ Procure and Deploy

      These are general operational steps, but not the specific steps of the AI Governance Improvement Model.

    • ✗ Code and Debug

      These are technical development activities rather than governance improvement activities.

    • ✗ Categorize and Route

      These are incident management activities, though they can be enabled by AI.

13. 13 Which 6C capability is primarily utilized when an AI agent automatically assigns an incident to the correct support team based on their current availability and skillset?

    This capability is about 'orchestrating' how work flows through the system.

    Coordination

    Coordination involves orchestrating workflows and the automated routing of tasks between parties.

    • ✗ Communication

      Communication is about the interface and message exchange, not the underlying routing logic.

    • ✗ Curation

      Curation filters data but does not necessarily trigger the operational workflow of assigning it to a person.

    • ✗ Creation

      Creation would involve drafting the ticket content, while Coordination handles the movement of that ticket.

14. 14 What does the 'Opacity' characteristic of AI refer to in governance discussions?

    Think about the challenges of 'Explainability'.

    The difficulty in explaining the internal logic of a model's output

    Opacity describes the 'black box' nature of complex AI, making it hard to audit the exact reasoning for a decision.

    • ✗ The inability of AI to work with unstructured data

      AI is actually often used specifically for unstructured data; opacity is about transparency of logic.

    • ✗ The physical security of the data center housing the AI

      Opacity is a logical and ethical characteristic of the software, not a physical security state.

    • ✗ The lack of user interface options for a model

      A model can have a clear interface but still have opaque internal reasoning.

15. 15 In the context of the EU AI Act mentioned in ITIL 5, AI systems affecting critical infrastructure are likely to be categorized as:

    These systems have the highest impact on safety and societal functioning.

    High-risk

    High-risk systems under the Act include those used in critical infrastructure due to the potential severity of failure.

    • ✗ Unacceptable risk

      Unacceptable risks are usually those that violate fundamental rights and are banned, rather than just high-stakes infrastructure.

    • ✗ Low-risk

      Critical infrastructure has high impact potential, making it the opposite of low-risk.

    • ✗ Non-regulated

      Critical infrastructure systems are a primary focus of AI regulatory frameworks.

16. 16 Which AI governance perspective focuses on implementing testing procedures to detect and mitigate algorithmic bias?

    It is about making sure the AI 'behaves' fairly and according to human values.

    Ethical Principles and Responsible AI

    This perspective embeds values like fairness into the testing lifecycle to catch biases before production.

    • ✗ Decision Authority and Risk Management

      This perspective deals with the delegation of power rather than the social fairness of the logic.

    • ✗ Regulatory Compliance

      Regulation may require these tests, but the operational practice of ensuring ethics belongs to the specific Ethical perspective.

    • ✗ Data Governance and Performance Management

      This deals with technical accuracy and drift, whereas bias is a sociocultural ethical risk.

17. 17 According to ITIL 5, why is traditional IT governance like COBIT or ITIL 4 insufficient for AI?

    Traditional systems do exactly what they are programmed to do, every time.

    It was designed for deterministic systems with predictable outputs

    AI is probabilistic and adaptive, whereas older frameworks assumed systems would follow fixed, predictable code paths.

    • ✗ It requires too much manual documentation

      AI governance actually increases documentation requirements; the issue is the 'nature' of the system behavior.

    • ✗ It focuses solely on hardware and infrastructure

      Traditional frameworks also focused on services and software, but they assumed those systems were static and predictable.

    • ✗ It is only applicable to on-premise environments

      Traditional governance adapted to the cloud; the issue is specifically the 'learning' and 'autonomous' nature of AI.

18. 18 What is the primary function of the 'Coordination' capability in the 6C model?

    It acts like a traffic controller for organizational processes.

    Automating routing decisions and orchestrating workflows

    Coordination handles the movement of work and the triggering of automated responses or assignments.

    • ✗ Summarizing log data for human review

      Summarizing is a function of Clarification or Creation, not workflow orchestration.

    • ✗ Identifying and removing duplicate data alerts

      This is Curation, which cleans and organizes data before it is routed.

    • ✗ Predicting resource demands for the next quarter

      Prediction and forecasting are functions of the Cognition capability.

19. 19 The 'Decision Authority and Risk Management' perspective is responsible for defining:

    It ensures that a person can still take control when the machine is wrong.

    Human escalation paths for automated decisions

    This perspective ensures there is a clear human path to override or handle anomalies in automated systems.

    • ✗ The toxicity score of Generative AI outputs

      Toxicity scoring is part of performance management and ethical monitoring.

    • ✗ The accuracy of training datasets

      Training data quality is the domain of Data Governance.

    • ✗ The alignment of AI with sustainability goals

      Sustainability is a broader strategic value rather than a specific decision authority control.

20. 20 Under ITIL 5's 'AI-native' philosophy, AI should be viewed as:

    It is no longer just a 'tool' on the side.

    An active partner in value streams

    ITIL 5 shifts from viewing AI as a tool to seeing it as a core participant in the product and service lifecycle.

    • ✗ A replacement for low-skilled service desk staff

      The framework emphasizes human-technology collaboration rather than simple replacement.

    • ✗ A separate infrastructure layer for IT only

      AI is integrated across the enterprise, including roles in HR, Finance, and Legal.

    • ✗ An optional extension for high-maturity teams

      ITIL 5 assumes AI is unavoidable and 'native' to the standard operating environment.

21. 21 Which 6C capability is most likely used to draft the initial version of a Service Level Agreement (SLA)?

    This capability involves 'producing' a new digital asset.

    Creation

    Creation involves generating new assets, reports, or drafts based on inputs.

    • ✗ Cognition

      Cognition might predict the needed targets, but the act of drafting the document is Creation.

    • ✗ Curation

      Curation organizes existing data rather than drafting new content.

    • ✗ Communication

      Communication is about the exchange with people, while Creation is about the production of the asset itself.

22. 22 What is 'Model Drift' in the context of ITIL 5 AI Governance?

    It is a performance issue that happens when the model gets 'out of sync' with reality.

    The degradation of model accuracy over time as it ingests new data

    Drift occurs when the model's performance shifts away from its intended parameters due to evolving data patterns.

    • ✗ The movement of a model from a local server to the cloud

      This is a deployment activity, not a governance risk like performance drift.

    • ✗ The change in ownership of a model between departments

      This is an organizational change; drift refers to the technical behavior of the model.

    • ✗ The process of a model learning a second language

      Learning is a feature; drift is specifically a risk associated with unintended changes in output quality.

23. 23 Which 6C capability helps organizations move from 'manual hand-offs' to 'high-velocity' value streams?

    This capability handles the orchestration of actions within a process.

    Coordination

    Coordination automates the flow of work, eliminating the delays found in traditional manual hand-offs.

    • ✗ Clarification

      Clarification explains the work, but Coordination is what actually moves it through the stream.

    • ✗ Curation

      Curation organizes the data being handed off, but it doesn't perform the 'hand-off' itself.

    • ✗ Cognition

      Cognition decides 'what' to do, but Coordination is the mechanism for 'doing' it at high velocity.

24. 24 The ITIL 5 AI Governance perspective on 'Regulatory Compliance' is uniquely challenging because:

    Think about how often AI laws and standards are changing right now.

    The legal landscape is evolving as rapidly as the technology

    Frameworks like the EU AI Act are new and frequently updated, requiring dynamic rather than static compliance.

    • ✗ Most AI systems are exempt from traditional laws

      AI is increasingly subject to specific and rigorous legal frameworks globally.

    • ✗ It requires lawyers to learn to write machine code

      Compliance is about oversight and policy alignment, not technical coding by legal staff.

    • ✗ It only applies to companies based in the United States

      Regulatory frameworks like the EU AI Act have global reach for any firm with European exposure.

25. 25 In the 'Assess' step of the ITIL AI Governance Improvement Model, what is meant by 'Stress-Testing'?

    It involves finding the limits of your 'rules' and 'oversight'.

    Identifying breaking points in current governance and performance

    Stress-testing evaluates where the existing governance might fail under pressure or at scale.

    • ✗ Running a model until the hardware overheats

      This is a physical hardware test; governance stress-testing is about policy and oversight limits.

    • ✗ Testing the AI with a massive amount of spam data

      While related to data robustness, governance stress-testing is specifically about the 'oversight' logic.

    • ✗ Simulating a complete data center power failure

      This is disaster recovery testing; AI governance focuses on the responsible use and decision-making logic.

26. 26 Which 6C capability is used by a virtual assistant to interact with a user in natural language to resolve a service request?

    This is the primary 'interface' capability for human stakeholders.

    Communication

    Communication involves the natural language interaction between humans and AI systems.

    • ✗ Clarification

      Clarification simplifies data, but Communication is the specific interface of exchange.

    • ✗ Coordination

      Coordination would fulfill the request in the background, but the assistant's dialogue is Communication.

    • ✗ Curation

      Curation might find the answer, but the act of 'talking' to the user is Communication.

27. 27 The ITIL 5 perspective 'Ethical Principles' includes 'Fairness.' How is this operationalized in governance?

    It involves checking if the 'math' treats everyone the same.

    By implementing testing to mitigate algorithmic bias

    Fairness is turned into an operational control by testing models for discriminatory patterns.

    • ✗ By ensuring the AI model is free to use for all employees

      This is an access policy; ethical fairness is about the quality and equity of the model's outcomes.

    • ✗ By providing a user manual for every automated tool

      Documentation is important, but fairness specifically addresses the bias and equity of the logic.

    • ✗ By setting a maximum price for AI-enabled services

      This is a commercial or financial decision, not a core ethical fairness control for model behavior.

28. 28 What is the primary risk of an 'Ungoverned Model Portfolio' according to ITIL 5?

    Consider the 'hidden' costs of taking risks that no one is monitoring.

    Liability that cannot be priced

    Without governance, the organization is exposed to risks (legal, reputational, financial) that it cannot accurately assess or control.

    • ✗ Slow deployment of new features

      Ungoverned portfolios often move 'too fast,' creating risk because they lack necessary gates and reviews.

    • ✗ A lack of computing power for training

      Computing power is a technical resource issue; governance is about the oversight and accountability of what is built.

    • ✗ Lower employee engagement with AI

      Employees often embrace ungoverned 'Shadow AI'; the risk is the liability, not the lack of use.

29. 29 Which 6C capability enables 'high-velocity' teams to automatically assign on-call personnel when a critical event is detected?

    This capability functions as the 'orchestrator' of human resources during an incident.

    Coordination

    Coordination automates routing and task assignment, which is essential for high-velocity incident response.

    • ✗ Creation

      Creation might draft the alert, but it doesn't handle the logistical assignment of a person.

    • ✗ Clarification

      Clarification helps the person understand the alert once they are assigned.

    • ✗ Cognition

      Cognition identifies that the alert is critical, but Coordination performs the act of assigning the responder.

30. 30 In ITIL 5, what does 'AI Governance Maturity' level 5 (Governance Advantage) represent?

    It is when 'doing the right thing' helps you win in the market.

    Governance is a trust signal and a competitive differentiator

    At the highest level, mature governance builds so much trust that it becomes a strategic advantage for the business.

    • ✗ All AI systems are fully automated with zero human oversight

      Full automation without oversight is actually a sign of poor governance, not high maturity.

    • ✗ The organization has the largest AI budget in its industry

      Budget is a measure of investment, while maturity is a measure of oversight, control, and value alignment.

    • ✗ Every employee is required to be certified in ITIL 5 Foundation

      Certifications are a capability builder, but maturity level 5 is about the 'strategic outcome' of the governance itself.