AIGP

Why do AI systems usually need governance controls beyond the controls used for ordinary deterministic software?

• AAI systems always run on public cloud infrastructure
• BAI outputs can be probabilistic, data-dependent, and difficult to predict in edge cases
• CAI systems do not interact with legal requirements
• DAI systems cannot be tested before release

Which characteristic of AI most directly increases the need for human oversight and clear accountability?

• AAI systems usually use smaller codebases than legacy software
• BAI systems can operate with opacity and at high speed or scale
• CAI systems require more electricity than spreadsheets
• DAI systems always replace all human decision makers

A hiring model is tested separately for different applicant groups to confirm that error rates are not materially worse for one protected group than another. Which responsible AI principle is being emphasized most directly?

• AFairness
• BData portability
• CRevenue optimization
• DOpen-source licensing

Which example best illustrates an AI harm that extends beyond a single individual and can affect society more broadly?

• AA model takes longer than expected to load in a lab environment
• BA public-facing generative system enables scalable deepfake impersonation during an elect…
• CA developer prefers Python to another language
• DA model card is formatted differently from the team template

Which statement best distinguishes a generative AI system from a traditional classification model?

• AA generative AI system can create new content such as text or images rather than only ass…
• BA generative AI system cannot use training data
• CA traditional classification model never makes mistakes
• DA generative AI system is always compliant if it is open source

What does a human-centric approach to AI governance emphasize most?

• AReplacing every human review step with full automation
• BDesigning and using AI in ways that respect human rights, agency, and meaningful oversight
• CPrioritizing model size over user impact
• DIgnoring user complaints until retraining is complete

A sales-forecasting model is optimized only for short-term revenue, and teams later discover that it repeatedly steers resources away from a strategic but lower-volume customer segment. Which risk is most directly illustrated?

• AMisalignment between model objectives and organizational goals
• BQuantum computing exposure
• CA complete lack of training data
• DA confidentiality breach by default

Why is data dependency a central governance concern for AI systems?

• ABecause AI systems only work with synthetic data
• BBecause the quality, relevance, and representativeness of data strongly affect system beh…
• CBecause data dependency eliminates the need for testing
• DBecause it guarantees explainability

A regulated lender wants to use a complex model for credit decisions. From a governance perspective, why is explainability especially important here?

• AIt guarantees the model will maximize approvals
• BIt helps the organization understand, justify, challenge, and monitor consequential decis…
• CIt removes all discrimination risk automatically
• DIt makes recordkeeping optional

Which control most directly supports the responsible AI principle of safety and reliability?

• ASkipping edge-case testing to meet the launch date
• BDefining performance thresholds and testing the system under realistic failure conditions
• CAllowing any employee to modify the production model
• DUsing only marketing claims to evaluate quality

Why do agentic AI systems often require stricter governance than a simple retrieval-only chatbot?

• AAgentic systems never rely on prompts
• BAgentic systems can take multi-step actions with more autonomy, which increases the need…
• CRetrieval-only chatbots are always illegal
• DAgentic systems cannot be limited by policy

Which action best reflects the responsible AI principle of accountability?

• AKeeping ownership informal so teams can move faster
• BAssigning named owners for approvals, monitoring, incidents, and remediation actions
• CAssuming the vendor is responsible for all downstream use
• DLetting the model decide when it should be audited

A company is forming its first AI governance program. What is the best initial structure for ongoing decision-making?

• AA cross-functional governance group with legal, privacy, security, product, technical, an…
• BA single engineer making all approvals alone
• CAn ad hoc chat group with no standing roles
• DA committee made only of external consultants

Which governance design choice most clearly establishes organizational expectations for AI use?

• ATelling teams to innovate first and document later
• BDefining approved use cases, prohibited uses, escalation paths, and required reviews
• CAllowing each team to create its own private definition of AI risk
• DRelying only on the vendor's website for internal guidance

Who should receive AI governance training in a mature program?

• AOnly the machine learning engineering team
• BAll relevant stakeholders, with depth tailored to their role and exposure
• COnly the board of directors
• DOnly external vendors

A global bank and a small regional retailer both deploy customer-service chatbots. Why should their AI governance programs differ?

• ABecause AI governance is optional for smaller organizations
• BBecause governance should be calibrated to size, maturity, industry, products, objectives…
• CBecause only banks face reputational risk
• DBecause retailers can ignore vendor management

Which statement best describes a deployer from an AI governance perspective?

• AThe party that uses an AI system in its own operational context
• BThe party that manufactured the servers
• CThe party that invented statistics
• DThe party that only reads media coverage about AI