Knowledge quiz

It's time to put what you've learned to the test, get 8 right to pass this unit.

What is the primary goal of implementing Responsible AI in an organization?

Choose the correct answer.

To limit AI adoption to low-risk use cases to avoid any potential business impact.

To ensure that AI systems are designed with transparency, accountability, and security to enable safe innovation.

To automate all business decisions and remove the need for human oversight in critical processes.

To focus exclusively on the efficiency gains of AI without considering regulatory or ethical requirements.

Which resource provides customers with subprocessor lists and Data Transfer Factsheets to ensure transparency in AI data processing?

Choose the correct answer.

SAP API Business Hub

SAP Trust Center

SAP Support Portal

SAP Community

In which AI scenario does processing occur directly within the SAP application system boundary?

Choose the correct answer.

Embedded AI scenarios

Side-by-side AI scenarios

Generative AI scenarios

External model provider scenarios

How does SAP ensure customer data protection when using external foundation models?

Choose the correct answer.

By allowing external providers to use customer data for retraining their models under strict supervision.

By enforcing contractual and technical controls that prevent external providers from using customer data for model training.

By requiring customers to manage all data masking and filtering before sending requests to SAP.

By storing all customer data in external model databases for faster processing of future requests.

What is the primary difference between training and inference in AI?

Choose the correct answer.

Training is the process of building or improving a model using data, while inference is when a trained model processes input to generate a result.

Training is when a model generates a response for a user, while inference is the process of building the model.

Training is an optional feature for external models, while inference is only used for internal SAP services.

Training involves processing data for a specific request, while inference involves storing data for long-term use.

How does SAP manage interactions with third-party AI model providers in generative AI use cases?

Choose the correct answer.

By allowing applications to connect directly to external provider APIs.

By routing requests through SAP-managed services such as the Generative AI Hub.

By delegating all governance and control to the third-party model providers.

By using open-source models that do not require SAP mediation.

Which environment provides European customers with a dedicated infrastructure for building and running AI in compliance with regional regulations?

Choose the correct answer.

SAP Business Technology Platform Global

EU AI Cloud

SAP API Business Hub

SAP Support Portal

Which statement best describes how interaction with AI systems differs from traditional software systems?

Choose the correct answer.

AI systems follow predefined rules and menus, while traditional software uses natural language processing.

AI systems interpret input and generate responses dynamically, whereas traditional software follows structured and predictable steps.

Traditional software requires constant retraining with new data, while AI systems remain static once they are deployed.

AI systems only retrieve existing information, while traditional software creates new content based on context.

How is data isolation enforced in SAP AI Core?

Choose the correct answer.

By enforcing separation at the infrastructure and runtime level

By allowing customers to share resource groups across tenants

By processing all customer data in a single shared environment

By requiring manual configuration of sandboxed environments for each workload

10.

How does SAP Sovereign Cloud support advanced sovereignty requirements?

Choose the correct answer.

By providing sovereignty across data, legal, operational, and technical dimensions.

By allowing customers to bypass all regional data protection laws.

By hosting all AI models on public, unmanaged infrastructure.

By removing the need for encryption and access controls.