3) Responsible Artificial Intelligence Practices (P1)

Question 2

What is responsible AI?

Answer 2

Responsible AI refers to practices and principles that ensure that AI systems are transparent and trustworthy while mitigating potential risks and negative outcomes.

Responsible AI emphasizes accountability and ethical considerations in AI development and deployment.

Question 3

What is the number one problem that developers face in AI applications?

Answer 3

Accuracy

Addressing issues of bias and variance is critical to improving accuracy in AI models.

Question 4

Define bias in the context of AI models.

Answer 4

Bias in a model means that the model is missing important features of the datasets, resulting in overly simplistic data representation.

Bias is assessed by the difference between expected predictions and true values.

Question 5

What does high bias indicate about a model?

Answer 5

When a model has high bias, it is underfitted, meaning it does not capture enough variation in the data features.

Underfitting leads to poor performance on training data.

Question 6

Define variance in the context of AI models.

Answer 6

Variance refers to the model’s sensitivity to fluctuations or noise in the training data.

High variance can lead to overfitting, where a model performs well on training data but poorly on new data.

Question 7

What is overfitting?

Answer 7

Overfitting occurs when a model performs well on training data but fails to generalize to unseen examples.

It happens when the model memorizes the training data rather than learning general patterns.

Question 8

List strategies to overcome bias and variance errors.

Answer 8

• Cross-validation
• Increase data
• Regularization
• Simpler models
• Dimension Reduction
• Stop training early

These strategies help improve model robustness and accuracy.

Question 9

What are some challenges of generative AI?

Answer 9

• Toxicity
• Hallucinations
• Intellectual property
• Plagiarism and cheating
• Disruption of the nature of the work

These challenges highlight ethical concerns and the impact of generative AI on society.

Question 10

What are the core dimensions of responsible AI?

Answer 10

• Fairness
• Explainability
• Privacy and security
• Veracity and robustness
• Governance
• Transparency
• Safety
• Controllability

Each dimension addresses specific ethical and operational aspects of AI systems.

Question 11

Define fairness in AI systems.

Answer 11

Fairness in AI systems promotes inclusion, prevents discrimination, upholds responsible values and legal norms, and builds trust with society.

Fairness is essential for the ethical deployment of AI technologies.

Question 12

What does explainability in AI refer to?

Answer 12

Explainability refers to the ability of an AI system to clearly explain or provide justification for its internal mechanisms and decisions.

This is crucial for user trust and understanding.

Question 13

What is meant by privacy and security in responsible AI?

Answer 13

Privacy and security ensure that users can trust their data is not compromised or used without authorization.

Protecting user data is a fundamental ethical requirement.

Question 14

What do veracity and robustness in AI involve?

Answer 14

Veracity and robustness refer to mechanisms that ensure an AI system operates reliably, even in unexpected situations, uncertainty, and errors.

These qualities contribute to the reliability of AI systems.

Question 15

What is governance in the context of responsible AI?

Answer 15

Governance is a set of processes used to define, implement, and enforce responsible AI practices within an organization.

Effective governance is essential for accountability in AI.

Question 16

Define transparency in responsible AI.

Answer 16

Transparency provides individuals, organizations, and stakeholders access to assess the fairness, robustness, and explainability of AI systems.

Transparency is crucial for building trust in AI technologies.

Question 17

What does safety in responsible AI refer to?

Answer 17

Safety refers to the development of algorithms, models, and systems that are responsible, safe, and beneficial for individuals and society.

Ensuring safety is a key aspect of ethical AI development.

Question 18

What is controllability in responsible AI?

Answer 18

Controllability refers to the ability to monitor and guide an AI system’s behavior to align with human values and intent.

This ensures that AI systems act in ways that are consistent with societal norms.

Question 19

List business benefits of responsible AI.

Answer 19

• Increased Trust and reputation
• Regulatory Compliance
• Mitigating Risks
• Competitive advantage
• Improved Decision Making
• Improved products and business

Implementing responsible AI can lead to significant strategic advantages for organizations.

Question 20

Model evaluation on Amazon Bedrock

Answer 20

you can evaluate, compare, and select the best foundation model for your use case,offers a choice of automatic evaluation and human evaluation.

Question 21

Model evaluation on SageMaker Clarify

Answer 21

You can automatically evaluate FMs for your generative AI use case with metrics such as accuracy, robustness, and toxicity to support your responsible AI initiative.

Question 22

Safeguards for generative AI

Answer 22

With Amazon Bedrock Guardrails, you can implement safeguards for your generative AI applications based on your use cases and responsible AI policies.

Question 23

Bias detection: SageMaker Clarify

Answer 23

helps identify potential bias in machine learning models and datasets without the need for extensive coding.
You specify input features, such as gender or age, and SageMaker Clarify runs an analysis job to detect potential bias in those features.

Question 24

Bias detection: SageMaker Data Wrangler

Answer 24

balance your data in cases of any imbalances.
Offers three balancing operators: random undersampling, random oversampling, and Synthetic Minority Oversampling Technique (SMOTE).

Question 25

Model prediction explanation: SageMaker Clarify

Answer 25

provide scores detailing which features contributed the most to your model prediction on a particular input for tabular, natural language processing (NLP), and computer vision models. For tabular datasets, SageMaker Clarify can also output an aggregated feature importance chart that provides insights into the overall prediction process of the model. These details can help determine if a particular model input has more influence than expected on overall model behavior.

Question 26

Monitoring and human reviews: SageMaker Model Monitor

Answer 26

monitors the quality of SageMaker machine learning models in production. You can set up continuous monitoring with a real-time endpoint (or a batch transform job that runs regularly), or on-schedule monitoring for asynchronous batch transform jobs.

Question 27

Monitoring and human reviews: Amazon Augmented AI (Amazon A2I)

Answer 27

service that helps build the workflows required for human review of ML predictions.

Question 28

Governance improvement: SageMaker Role Manager

Answer 28

With SageMaker Role Manager, administrators can define minimum permissions in minutes.

Question 29

Governance improvement: SageMaker Model Cards

Answer 29

With SageMaker Model Cards, you can capture, retrieve, and share essential model information, such as intended uses, risk ratings, and training details, from conception to deployment.

Question 30

Governance improvement: SageMaker Model Dashboard

Answer 30

With SageMaker Model Dashboard, you can keep your team informed on model behavior in production, all in one place.

Question 31

AWS AI Service Cards:

Answer 31

resource to help you better understand AWS AI services. AI Service Cards are a form of responsible AI documentation that provides a single place to find information on the intended use cases and limitations, responsible AI design choices, and deployment and performance optimization best practices for AWS AI services.

Question 32

Components of an AWS AI Service Cards:

Answer 32

Basic concepts to help customers better understand the service or service features Intended use cases and limitations Responsible AI design considerations Guidance on deployment and performance optimization

Question 33

Responsible Considerations to Select a Model

Answer 33

Define application use case narrowly Choosing a model based on performance Choosing a model based on sustainability concerns

Question 34

Responsible Preparation for Datasets : Balancing datasets

Answer 34

Balanced datasets are important for creating responsible AI models that do not unfairly discriminate or exhibit unwanted biases.

Question 35

Responsible Preparation for Datasets: Inclusive and diverse data collection

Answer 35

Inclusiveness and diversity in data collection ensure that data collection processes are fair and unbiased. Data collection should accurately reflect the diverse perspectives and experiences required for the use case of the AI system. This includes a diverse range of sources, viewpoints, and demographics. By doing this, the AI system can work to ensure decisions are unbiased in their performance.

Question 36

Responsible Preparation for Datasets: Data curation:

Answer 36

Curating datasets is the process of labeling, organizing, and preprocessing the data so that it can perform accurately on the model.