AI-900 Question 69
Select 3You are designing an AI-powered loan approval system for a financial institution. The institution wants to ensure the AI solution is transparent to build trust with customers and comply with regulatory requirements. Which considerations should you prioritize to ensure transparency in the AI solution?
- A
Ensure the model's decision-making process can be explained to non-technical stakeholders.
- B
Limit access to the AI system's decision-making logic for security reasons.
- C
Document the data sources and any preprocessing steps used in the model.
- D
Provide a clear explanation to users of how their data influences AI decisions.
- E
Focus solely on achieving the highest accuracy for predictions, regardless of interpretability.
Show answer and explanation
Correct answers: A, C, D
Explanation
To ensure transparency in an AI solution, it is important to provide explainability, document processes, and maintain clear communication with users about how the system functions and uses their data. Transparency fosters trust and compliance with legal and ethical standards, particularly in sensitive applications like loan approvals.
- A. Correct.
Ensuring the model's decision-making process can be explained to non-technical stakeholders is crucial for transparency, as it helps build trust and promotes understanding of the system's outputs.
- B. Incorrect.
Limiting access to the decision-making logic may hinder transparency, as stakeholders might not be able to understand or validate the AI system's outputs.
- C. Correct.
Documenting data sources and preprocessing steps is essential for transparency, as it allows stakeholders to understand the origins and transformations of the data used in the model.
- D. Correct.
Providing a clear explanation of how user data influences AI decisions is a key aspect of transparency, ensuring users understand the role their data plays in the system's outputs.
- E. Incorrect.
Focusing solely on accuracy without considering interpretability undermines transparency, as stakeholders and users may not understand how decisions are made, leading to mistrust.