AI as a Service CT-AI Practice Questions (Page 3)

Questions 11–15

1. 11expert · hard

   A law firm is considering an AIaaS provider's contract-analysis API. The firm requires that the provider's employees cannot access the documents and that the provider cannot use the documents to improve its models. The provider offers a standard contract that includes a data processing agreement but does not explicitly restrict the provider's use of the data for model improvement. The firm's IT team suggests negotiating an addendum. What should the firm do?

   • ABuild an in-house contract-analysis tool instead of using AIaaS.
   • BUse the API but only for non-sensitive documents.
   • CNegotiate an addendum that explicitly prohibits the provider from using the firm's documents for model improvement.
   • DAccept the standard contract because the data processing agreement is sufficient.
   Reveal Solution Discussion Select an answer first
2. 12expert · hard

   A company is using an AIaaS provider's predictive-analytics API. The API has an SLA that guarantees 99.9% uptime. However, the company experiences a service outage that lasts 30 minutes, and the API returns incorrect predictions for several hours after the outage. The company wants to claim compensation for the incorrect predictions. Which contract element should the company review?

   • AThe SLA's uptime guarantee, because the outage exceeded the allowed downtime.
   • BThe data ownership terms, because they define who owns the predictions.
   • CThe liability clause, because it defines the provider's responsibility for damages caused by incorrect outputs.
   • DThe pricing model, because the company should not pay for the outage period.
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A media company is using an AIaaS provider's content-moderation API. The company wants to ensure that the provider does not use the content to improve the provider's models for other customers. Which contract element should the company review?

   • AThe pricing model that charges per image analyzed.
   • BThe data usage terms that specify whether customer data can be used for model training.
   • CThe service level agreement (SLA) that defines the API's uptime.
   • DThe liability clause that limits the provider's responsibility for missed content.
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A non-profit organization wants to add a sentiment-analysis feature to its survey tool. They have a limited budget and no IT staff. They need the feature to be available quickly and to handle occasional spikes in usage. Which approach is most consistent with the AIaaS model?

   • AManually review survey responses and classify sentiment by hand.
   • BTrain a custom sentiment model using a cloud-based machine learning platform.
   • CUse a pre-trained sentiment-analysis API from a cloud provider.
   • DInstall an open-source sentiment-analysis library on a rented server.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A startup wants to add speech-to-text transcription to its meeting-recording app. The team has no machine-learning expertise and wants to minimize development time. They expect usage to grow rapidly. Which AIaaS offering is most appropriate?

   • AA chatbot service that can transcribe audio messages.
   • BA cloud-based machine learning platform to train a custom speech model.
   • CAn on-premises speech recognition server that the startup purchases and maintains.
   • DA pre-trained speech-to-text API from a cloud provider.
   Reveal Solution Discussion Select an answer first