1What factor of trust asks if the system can be perceived or touched?
Tangibility
2What factor of trust asks if the underlying operating rules and inner logics are apparent to users?
Transparency (explainable AI)
3What factor of trust asks if the system exhibits the same and expected behavior over time?
Reliability
4What factor of trust relates to the actual work the system is performing (e.g., technical vs. interpersonal judgments)?
Task characteristics
5What factor of trust relates to the degree of non-verbal, physical, or psychological closeness performed by the system?
Immediacy behaviors
6What is an example of a 'local explanation' method in XAI?
LIME (Local Interpretable Model-Agnostic Explanations)
7What is an example of a 'global explanation' method in XAI?
SHAP (SHapley Additive exPlanations)
8What is "Google's style" of de-biasing?
Identify the decision point and de-bias the actual decision, regardless of what the system learned from unfair data.
9What is "Microsoft's style" of de-biasing?
Change the biased data until it's more fair before training the model.
10How do humans naturally perceive interacting agents, including AI?
By attributing intentions, beliefs, and emotions (Theory of Mind).
11When do people tend to trust AI more?
For objective, rule-based tasks (vs. subjective, intuitive ones).
12How might trust dynamics differ between robotic and virtual AI?
Trust in robotic AI may start low and increase, while trust in virtual AI might start high and decrease upon errors.
13What is 'Algorithm Aversion'?
When trust in an AI decreases after observing its limitations or errors, often starting from a high initial expectation.
14What is the 'Uncanny Valley'?
The dip in familiarity or trust when humanoid robots look almost perfectly human but have subtle flaws.
15What is a major challenge for Explainable AI (XAI)?
AI decision processes can be complex and unlike human reasoning (and humans aren't always rational explainers either).
16What are 'Intrinsically Interpretable Models'?
Using simpler models (like logistic regression) that are easier to understand by nature.
17What are 'Post-hoc Explanations'?
Methods used to explain existing complex models (like neural networks) after they are trained.
18How does LIME generate a local explanation?
It creates perturbations (small changes) of an instance, gets predictions from the complex model, and fits a simple local model to explain that one prediction.
19What is the 'Dual Use' problem in AI?
The fact that AI technologies can be used for both beneficial and harmful purposes (e.g., security vs. surveillance).
20What is the 'Moral Machine' project related to?
Investigating human perspectives on moral dilemmas for AI, such as self-driving car crash scenarios.
21Why is "Fairness through blindness" (removing sensitive attributes) often insufficient?
Because other data features are often correlated with the sensitive attributes anyway.