Computer Ethics

Why do people find AI models untrustworthy due to the "black box problem"?

• AI models are too expensive to use

• AI decisions are often not easily explainable or transparent

• AI models never make mistakes

• AI models always require human oversight

How can you ensure that an AI model is fair for everyone?

• By avoiding testing AI models before deployment

• By making AI models completely random in decision-making

• By allowing AI to learn without any human oversight

• By using diverse and representative training data

Why is it important to explain what dataset is used to train an AI model?

• It makes the AI model work faster

• It helps people understand potential biases and limitations in the AI model

• It prevents the AI model from making mistakes

• It ensures the AI model never needs updates

Why is it important for AI to be designed without benefiting one person over another?

• ) To ensure only certain groups get the best results from AI

• It ensures AI models are independently tested for reliability and fairness

• o ensure fairness and prevent discrimination in decision-making

• It allows AI to operate without any regulations

Why is understanding the metadata of an AI model important?

• It helps track the data sources, performance, and potential biases of the model

• It makes the AI model faster and more efficient

• It allows the AI to operate autonomously without supervision

• It ensures that the model is completely error-free

Are you enjoying the video lesson?

• Yes

• No

Why is data privacy important when training an AI model?

• It helps AI models to make biased decisions based on private information

• It allows AI models to work faster by using any data

• It ensures personal data is protected from misuse and unauthorised access

• It ensures that AI models can operate without any regulations

What is one of the key social challenges of AI?

• AI models can replace all human jobs, leading to unemployment

• AI systems can be biased, leading to unfair outcomes and societal inequality

• AI makes all decisions purely based on logic and emotion

• AI models do not require any human input or data

What types of tools are needed to train an AI model correctly?

• A single algorithm that works for all types of AI models

• Random data sources with no validation process

• Only a powerful computer with no need for data preprocessing

• High-quality datasets, bias detection tools, and model evaluation frameworks

How do people and culture impact AI models?

• AI models are completely independent of human and cultural influence

• They influence biases in training data, ethical considerations, and trust in AI systems

• AI models function the same way in every culture without any differences

• People and culture have no impact on AI because it is purely mathematical