Generative AI Leader

Last checked October 22, 2026 against the official Google Cloud exam page.

1. Question 1Business strategies for a successful gen AI solution

   Which responsible AI principle is most directly concerned with ensuring model outputs and logs do not expose personal information?

   1. APrivacy and data protection for personal information
   2. BInterpretability, so each decision can be explained to a reviewer
   3. CLatency and throughput of the model endpoint under peak load
   4. DFairness, so that output quality is consistent across user groups
   Show answer and explanationHide answer

   Correct answer: A

   Privacy and data protection cover how personal data is collected, used, logged and exposed. Fairness concerns differential outcomes, interpretability concerns explaining decisions, and latency concerns performance, so none of them addresses personal data exposure directly.

2. Question 2Fundamentals of gen AI

   A model's answer about a company's refund policy is fluent and confident, but the policy it describes does not exist. Which term describes this failure?

   1. AOverfitting, where the model memorises its training examples
   2. BHallucination, where fluent output is factually wrong
   3. CTokenisation drift, where token counts change between model versions
   4. DLatent bias, where the model favours some topics over others
   Show answer and explanationHide answer

   Correct answer: B

   Hallucination is the term for fluent output that is presented as fact but is not supported by the data or the world. Overfitting describes a training problem rather than a confident invented answer, and tokenisation drift concerns how text is split, not whether it is true. Latent bias is about systematic skew, which is a different risk.

3. Question 3Techniques to improve gen AI model output

   Which statement about prompt design is most accurate?

   1. ALonger prompts always improve accuracy, so the prompt length should be maximised
   2. BPrompts must be written in English for the model to respond correctly at all
   3. CClear instructions, context and an example often improve output quality
   4. DPrompts cannot influence the output once a model is deployed to production
   Show answer and explanationHide answer

   Correct answer: C

   Clear instructions, relevant context and a worked example usually make outputs more accurate and better formatted. Length alone does not help, and many models handle several languages. Prompts remain one of the main levers after deployment.

4. Question 4Google Cloud's gen AI offerings

   Which metric best describes the usage cost of an LLM call on a pay-per-use service?

   1. AInput and output tokens processed by each request
   2. BThe total storage size of the model file in the cloud bucket
   3. CThe number of pages in the source document uploaded to the chat
   4. DThe number of employees who use the assistant in a given month
   Show answer and explanationHide answer

   Correct answer: A

   Pay-per-use LLM services generally bill for the tokens sent in and generated out by each request, so token volume is the cost driver to track. User counts, page counts and model file size may affect other things but are not what is billed per call.

5. Question 5Business strategies for a successful gen AI solution

   What is the main purpose of a safety filter on a generative AI endpoint?

   1. ATo shorten responses so they display well on mobile devices
   2. BTo compress the model weights so that they use less storage
   3. CTo speed up token generation for high-traffic endpoints
   4. DTo block or flag harmful or disallowed content
   Show answer and explanationHide answer

   Correct answer: D

   Safety filters screen inputs and outputs against policy categories and block or flag content that is harmful or disallowed. Response length, speed and weight compression are performance or formatting concerns and are not what a safety filter is for.