AI Assistants Quiz

Answers and explanations appear as you take the quiz.

1. What is an AI assistant in the guide’s definition?
   • An interface that helps a person get information, create material, or complete tasks
   • A model with unlimited access to every account
   • A database of verified answers
   • A chatbot that cannot use tools
2. If an assistant answers from a supplied document, what is its evidence basis?
   • The document automatically retrains the model
   • The model has verified every claim in the document
   • The provided document is part of its context, but the answer still needs checking
   • The assistant must ignore the supplied text
3. What does a saved preference in an assistant’s memory imply about training?
   • That every model parameter changed
   • Not necessarily that the underlying model was retrained
   • That the preference is public
   • That the assistant cannot forget anything
4. What determines whether an assistant can access a connected account?
   • The model’s name alone
   • A prompt asking it to ignore permissions
   • The amount of text in the conversation
   • The product integration and permissions granted
5. The assistant says it drafted a report but has not saved a file. How should it label the result?
   • As a draft, not a completed file
   • As a saved file because text exists
   • As an approved report
   • As a verified result
6. What interaction feature supports recoverability?
   • Forcing a restart after every error
   • Hiding whether an action is running
   • Allowing correction, cancellation, or manual workflow
   • Executing consequential actions without review
7. Why should a product explain whether an answer used supplied sources or tools?
   • It guarantees the answer is correct
   • The answer’s evidence basis affects how it should be checked
   • It determines the user’s account password
   • It removes the need for source links
8. Which metric better reflects assistant usefulness than answer length alone?
   • Number of sentences
   • Model size
   • Conversation duration
   • Task outcome and user effort