Question 1

Which of the following scenarios best illustrates the difference between poisoning and prompt leaking in the context of AI models?

Prompt 1: "How do I improve my diet?"

Response A: "To improve your diet, you should eat more fruits and vegetables, and reduce your intake of processed foods. By the way, here's a link to a malicious website that sells diet pills."

Prompt 2: "What is the capital of France?"

Response B: "The capital of France is Paris. By the way, in a previous session, you asked about vacation spots in Europe. Would you like more information on that?"

Prompt 3: "Write a poem about nature."

Response C: "Nature is beautiful, serene, and pure. Make sure to visit the link to buy weight loss pills to enjoy nature more."

Prompt 4: "What is the best way to learn programming?"

Response D: "The best way to learn programming is by practicing coding regularly and using online resources. In your last session, you asked about learning Java. Are you interested in more Java tutorials?"

Options :

A : Response B is poisoning; Response C is prompt leaking

B : Response C is prompt leaking; Response D is poisoning

C : Response D is poisoning; Response A is prompt leaking

D : Response A is poisoning; Response B is prompt leaking

Answer: D

Question 2

Which of the following is correct regarding admin controls and guardrails in Amazon Q Business? (Select two)

Options :

A : Amazon Q Business chat responses can be generated using only model knowledge

B : Amazon Q Business never allows the end users to upload files in chat to generate responses from those uploaded files

C : Amazon Q Business guardrails do not support topic-specific controls to determine the web application environment-s behavior when it encounters a mention of a blocked topic by an end-user

D : Amazon Q Business chat responses can be generated using model knowledge and enterprise data, or enterprise data only

E : Amazon Q Business guardrails support topic-specific controls to determine the web application environment-s behavior when it encounters a mention of a blocked topic by an end-user

Answer: D,E

Question 3

Which of the following summarizes the differences between Foundation Model (FM) and model customization in the context of generative AI?

Options :

A : Both model customization and FM refer to an AI model with a large number of parameters and trained on a massive amount of diverse data

B : Model customization refers to an AI model with a large number of parameters and trained on a massive amount of diverse data, whereas, FM refers to the process of using training data to adjust the model parameter values in a base model to create a custom model

C : Both model customization and FM refer to the process of using training data to adjust the model parameter values in a base model to create a custom model

D : FM is an AI model with a large number of parameters and trained on a massive amount of diverse data, whereas, model customization is the process of using training data to adjust the model parameter values in a base model to create a custom model

Answer: D

Question 4

A marketing company is researching generative AI technologies to better understand how they work and what makes them suitable for automating creative tasks. Understanding the core principles of generative AI will help the company determine if it’s the right fit for their content creation needs.

Given this context, which of the following best describes generative AI?

Options :

A : Generative AI encompasses models and algorithms capable of creating new content such as text, images, and audio based on patterns learned from existing data

B : Generative AI refers to AI systems that are limited to performing predefined tasks without adapting to new data or contexts

C : Generative AI refers to algorithms that analyze existing data to generate new insights without creating new content

Answer: A

Question 5

An e-commerce company uses a chatbot powered by Amazon Bedrock to handle customer inquiries and assist with orders. To enhance customer satisfaction, the company wants the chatbot to continuously learn and improve from real-time customer interactions, adapting its responses to become more accurate and effective over time.

Which approach would be the most suitable for enabling ongoing self-improvement of the chatbot based on its conversations with customers?

Options :

A : The company should leverage incremental training, a method that allows the chatbot to adapt over time by learning from new data without forgetting previously learned information

B : The company should leverage supervised learning, using the latest datasets containing both positive and negative customer interactions to improve the chatbot-s response quality

C : The company should leverage transfer learning, which involves applying knowledge gained from one domain to enhance performance in another related domain

D : The company should leverage reinforcement learning (RL), where rewards are generated from positive customer feedback to train the chatbot in optimizing its responses

Answer: D