AI-900 Practice Question: Describe features of generative AI workloads on Azure

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Automatically compiling source code into executable binaries

  Why it's wrong here

  Code compilation is a software build step, whereas code generation creates new source code from natural language descriptions. Compilation is a fully deterministic, rule-based translation of existing high-level source code into executable binaries or bytecode, driven by the compiler's grammar and semantic rules. It preserves program semantics exactly and never invents new logic, nor does it interpret human intent in a string of English words. In contrast, code generation is a probabilistic, model-driven synthesis that converts an unstructured prompt into novel source code—an act of creative derivation, not mechanical transformation.

• ✓

  AI producing programming code from natural language descriptions — used in IDEs and developer tools

  Why this is correct

  AI producing programming code from natural language descriptions is the essence of generative code models such as GitHub Copilot. These large language models are fine-tuned on huge code corpora and take a developer's comment or prompt—for example, 'sort this list ascending'—to generate syntactically and semantically correct code in the current editor context. They are embedded in IDEs to deliver real-time completions, whole-function generation, and even test suggestions, allowing developers to express intent in English and receive working source code. This directly matches the core definition of code generation as natural-language-to-code synthesis.

• ✗

  Scanning existing code for security vulnerabilities and generating a fix automatically

  Why it's wrong here

  Automated vulnerability patching is a specialised security AI use, whereas code generation broadly creates new code from natural language. Scanning existing code for security flaws relies on static and dynamic analysis tools that detect known vulnerability signatures and then propose or apply a minimal, localized patch. This is a reactive, repair-focused workflow that targets defects in an already-written codebase, not a creative synthesis of new logic from a plain-English description. The output and intent are fundamentally different: one fixes a flaw, the other produces original source code.

• ✗

  Auto-generating boilerplate project structure files when creating a new repository

  Why it's wrong here

  Project scaffolding tools generate templates, while code generation AI creates meaningful code logic from natural language specifications. Scaffolding a new repository is a deterministic template-expansion process that lays out predefined folders, configuration files, and placeholder source files without interpreting any business logic or intended behavior. It follows fixed blueprints and naming conventions, requiring no semantic inference or learning from data. Code generation, by contrast, parses a natural language prompt and synthesizes original, functional algorithms that do not exist in any template.

Go deeper

Related to this question