AWS Certified AI Practitioner AIF-C01 Real Mock 4

1. A risk management committee at a commercial bank audits an automated credit-scoring model. The audit reveals that the model rejects mortgage applications from specific demographic postal codes at a disproportionately higher rate, despite those applicants possessing equivalent income and debt-to-income ratios as approved applicants. Which ethical AI pillar is the bank violating, and which concept must they actively address to ensure equitable lending?

A. Model Fit

B. Fairness

C. Model Latency

2. A cybersecurity firm develops an automated defense bot. The bot continuously monitors network traffic, independently decides when a server behavior looks anomalous, and proactively alters firewall rules to isolate the threat to achieve its programmed goal of securing the network. Which AI concept best describes this adaptive, goal-oriented behavior?

A. Generative AI

B. Agentic AI

C. Deep Learning

3. A media company allows users to upload high-resolution 4K video files for automated AI-based subtitling. The input payloads are massive (often up to 1 GB), and the model takes up to 10 minutes to process each video. The client app submits the video and retrieves the results via a notification later. Which Amazon SageMaker inferencing strategy is required?

A. Real-time inferencing

B. Batch inferencing

C. Serverless inferencing

4. A HealthTech company is developing a diagnostic application using Amazon Rekognition Custom Labels. The application ingests massive datasets composed entirely of digital X-ray scans to identify microscopic bone fractures. What type of data is primarily being processed in this scenario?

A. Unstructured image data

B. Structured tabular data

C. Time-series sequence data