Question 6
Domain 5: Security, Compliance, and Governance for AI SolutionsA company has terabytes of data in a database that the company can use for business analysis. The company wants to build an AI-based application that can build a SQL query from input text that employees provide. The employees have minimal experience with technology. Which solution meets these requirements?
Correct answer: A
Explanation
GPT fits because it is a generative language model that can turn “input text” into structured output like SQL, which matches the need to build a query from natural language. The source describes Bedrock’s flow as “Invoke model” and “Response generated,” showing the model is used to generate the needed text output for non-technical users.
Why each option is right or wrong
A. Generative pre-trained transformers (GPT)
Amazon Bedrock Guardrails is the feature that enforces content policies at inference time, and it supports denied topics to block investment-product recommendations as well as sensitive information filters to detect and redact PII such as account numbers before the prompt is processed. The Bedrock request flow shown here explicitly includes “Submit with guardrails” before model invocation and an “Apply output filter” step afterward, matching the requirement to control both inbound user text and generated output in a production path.
B. Residual neural network
Residual networks are deep learning models mainly used for image and other structured pattern tasks.
C. Support vector machine
Support vector machines classify or regress on labeled features; they do not generate SQL from text.
D. WaveNet
WaveNet generates sequential audio waveforms, not natural-language database queries.