Managing Artificial Intelligence GSLC Practice Questions (Page 10)

Questions 46–50

1. 46application · medium

   A company's AI-based recommendation system is providing inappropriate content to users, causing reputational damage. The incident was detected by user complaints. The company has an incident response plan for traditional IT incidents but not for AI-specific incidents. Which action is most important to improve future response?

   • ADevelop an AI-specific incident response plan that includes model rollback, data preservation, and communication procedures.
   • BDisable the recommendation system permanently to avoid further damage.
   • CRely on the existing IT incident response plan, since AI incidents are similar to other IT incidents.
   • DIncrease the frequency of model retraining to prevent future incidents.
   Reveal Solution Discussion Select an answer first
2. 47foundation · easy

   Why should AI initiatives be aligned with an organization's risk appetite?

   • ATo ensure that AI projects do not expose the organization to unacceptable risks
   • BTo maximize the number of AI projects regardless of risk
   • CTo eliminate all risks from AI systems
   • DTo ensure AI projects are always the cheapest option
   Reveal Solution Discussion Select an answer first
4. 48application · medium

   A manufacturing company is introducing AI for predictive maintenance on factory equipment. The IT team has no prior experience with AI. The company wants to ensure that staff can manage the AI system effectively. Which action is most important for building the necessary skills?

   • ASend all IT staff to a week-long general AI conference to learn about the latest trends.
   • BPurchase a commercial AI maintenance tool and rely on the vendor's support for all issues.
   • CProvide targeted training for IT staff on AI fundamentals, model lifecycle management, and incident response.
   • DHire a team of data scientists to handle all AI-related tasks, and keep the IT staff focused on traditional systems.
   Reveal Solution Discussion Select an answer first
5. 49application · medium

   An organization deploys a chatbot that uses natural language processing to answer customer queries. Users discover they can trick the chatbot into revealing sensitive information by using specially crafted prompts. This is an example of which type of AI risk?

   • AData poisoning attack
   • BPrompt injection attack
   • CModel inversion attack
   • DDenial of service attack
   Reveal Solution Discussion Select an answer first
6. 50foundation · easy

   Which of the following is an AI-specific incident that might require a unique response?

   • AA server overheating
   • BA network cable failure
   • CA model producing biased or harmful outputs in production
   • DA user forgetting their password
   Reveal Solution Discussion Select an answer first