AI for Technical Writers Quiz

Answers and explanations appear as you take the quiz.

1. What does the Vale linter do in a docs-as-code workflow?
   • Generates API reference pages from code
   • Hosts the documentation website
   • Checks prose against style-guide rules and gives the same result every time
   • Translates docs into other languages
2. What best describes docs-as-code?
   • Documentation kept in version control and reviewed and built like code
   • Writing documentation only inside code comments
   • Using AI to write all docs automatically
   • Publishing documentation as PDF only
3. Which four content types does the Diátaxis framework separate?
   • Blogs, emails, slides and videos
   • Tutorials, how-to guides, reference and explanation
   • Drafts, reviews, approvals and archives
   • Beginner, intermediate, advanced and expert
4. Why should AI-generated code samples be run as tests?
   • Tests make pages load faster
   • Search engines require tested code
   • Testing replaces the need for a style guide
   • Models can invent plausible parameters or flags, and tests catch them before readers do
5. What is the llms.txt proposal?
   • A suggested file that points language models to a site's key documentation
   • A legal license for AI training
   • A style guide published by Microsoft
   • A required format for all API specs
6. Which writing practice helps AI assistants that retrieve documentation?
   • Long pages with no headings
   • Frequent references like 'as mentioned above'
   • Self-contained sections with descriptive headings and explicit product names
   • Putting all content in images
7. How was API reference documentation commonly produced before generative AI?
   • It was always written entirely by hand
   • It was generated from specs or code comments with tools such as Swagger UI, Redoc or Sphinx autodoc
   • It did not exist
   • It was produced by speech recognition
8. According to the guide, where does a technical writer add the most value alongside AI?
   • Typing faster than the model
   • Avoiding version control
   • Removing all human review
   • Checking accuracy, deciding what users need and structuring information