Google Generative-AI-Leader - Google Cloud Certified - Generative AI Leader Exam

Training a facial-recognition system predominantly on one demographic creates representation bias. The model receives insufficient examples from other groups and will probably perform less accurately for those populations, producing systematically unequal outcomes. This is directly associated with fairness because model performance varies according to demographic characteristics. The other scenarios describe serious data-quality problems, but their primary effects differ. Missing merchant details can reduce fraud-detection accuracy, incorrectly encoded text introduces corruption, and bot traffic distorts customer-behavior signals. Those defects may degrade overall performance without necessarily disadvantaging a protected or underrepresented group. Responsible AI development requires representative datasets, subgroup-level evaluation, documented data provenance, and ongoing monitoring for unequal error rates. Therefore, the facial-recognition dataset presents the clearest and most direct risk of an unfair and biased outcome.

The core requirement is to guarantee that the chatbot only uses information from the company ' s official documentation and does not rely on its general knowledge base. This is crucial for ensuring factual accuracy, relevance to the company ' s specific products, and preventing the generation of fabricated or incorrect information (hallucinations).

The specific technique designed to address this challenge is Grounding. Grounding is the process of connecting the Large Language Model ' s (LLM ' s) responses to a trusted, verifiable source of information, such as an organization ' s internal documents, databases, or live data feeds. When an LLM is grounded, it is forced to base its answers only on the provided context, effectively preventing it from drawing on its broad, generalized training data. Grounding is often implemented using a method called Retrieval-Augmented Generation (RAG), particularly with tools like Google Cloud ' s Vertex AI Search, which indexes the official documentation and feeds the relevant snippets to the model.

Options A, B, and C address different aspects of model output: Role prompting sets the model ' s persona, adjusting temperature controls creativity, and prompt chaining manages conversation history, but none of these techniques restrict the model ' s source of truth to the official documentation. Therefore, Grounding is the correct and most effective technique for this requirement.

===========

Multimodal search directly enhances the customer experience by allowing them to find products using various intuitive methods (images, voice, text). This leads to easier product discovery, higher engagement, and ultimately increased customer satisfaction and potential sales, which is a primary business benefit.

________________________________________

First-party information is data owned or directly collected by an organization, such as customer transactions, account records, support history, and purchase details. A chatbot cannot reliably answer a question about a customer’s recent purchase history from a foundation model’s general training data. It must be grounded in the company’s current, authorized customer records to provide an accurate and personalized response. Appropriate identity verification and access controls must also be applied before retrieving the information. Public sentiment is generally evaluated using external public data, while definitions of common scientific terms can normally be answered from general model knowledge. Summarizing global news requires grounding in external news sources rather than proprietary first-party records. Therefore, retrieving specific purchase-history information is the clearest situation where first-party grounding provides essential factual context.