AIncorrect
Fairness
Fairness is the characteristic of preventing specific groups from suffering unfair disadvantage.
This question is about people being able to control and stop the behavior, not fair treatment between groups, so this is incorrect.
A company places importance on having people oversee an AI system's behavior so that a person can intervene and stop it when undesirable behavior occurs. Which responsible AI characteristic BEST matches this idea?
1 / 1
Select an answer
CorrectD
Explanation
A question that asks which responsible AI characteristic lets people control and stop behavior.
- 1「people oversee」Maintain human oversight rather than leaving it to the AI
- 2「a person can intervene and stop it」Able to intervene and stop when needed = controllability
BIncorrect
Explainability
Explainability is the characteristic of being able to understand why an individual prediction was reached.
Even if you understand the rationale, that alone cannot stop undesirable behavior on the spot. This question is about people being able to oversee, intervene, and stop the system, which is controllability, so this is incorrect.
CIncorrect
Inclusivity
Inclusivity is the characteristic of addressing a wide range of diverse users and situations without leaving anyone out.
This question is about people being able to control and stop the behavior, not the breadth of coverage, so this is incorrect.
DCorrect
Controllability
Correct. Controllability is the responsible AI characteristic of having people oversee the AI's behavior and intervene or stop it as needed. It is the idea of maintaining human oversight.
Key Takeaway
Having people oversee behavior and intervene or stop it is controllability. It is distinct from fairness (between groups), transparency (disclosure), and inclusivity (breadth of coverage). It maintains human oversight rather than leaving everything to automation, and services such as Amazon A2I (human review) are related.