Explainability: Quiz

In the context of AI ethics, how is “interpretability” distinguished from “transparency”?

A) Interpretability is the legal right to sue a company, while transparency is the technical source code.

Correct Incorrect Correct answer

B) Interpretability focuses on what a user needs to change at the input to get a different result, while transparency focuses on the internal mechanics of the system.

Correct Incorrect Correct answer

C) Transparency is only for government regulators, while interpretability is only for AI researchers.

Correct Incorrect Correct answer

D) There is no distinction; both terms describe the process of making software open-source.

Correct Incorrect Correct answer

What is the primary difference between the industrial “many hands problem” and the AI “accountability gap”?

A) The many hands problem is a practical difficulty that can eventually be unraveled, whereas the accountability gap can be a theoretical impossibility due to the "black box" nature of AI.

Correct Incorrect Correct answer

B) The many hands problem is about physical parts, while the accountability gap is about digital electricity.

Correct Incorrect Correct answer

C) The accountability gap is easy to solve with better logging, while the many hands problem is legally impossible to fix.

Correct Incorrect Correct answer

D) The many hands problem only applies to manual labor, while the accountability gap only applies to software engineers.

Correct Incorrect Correct answer

According to the “performance-first” argument, why should we accept high-performing AI even if it is not explainable?

A) Because humans have a long history of using effective technology, like aspirin, without fully understanding how it works.

Correct Incorrect Correct answer

B) Because explainability is a secret method used by competitors to steal intellectual property.

Correct Incorrect Correct answer

C) Because the "right to explanation" is a modern invention that has no basis in human law or psychology.

Correct Incorrect Correct answer

D) Because machines are inherently more moral than humans and do not require oversight.

Correct Incorrect Correct answer

How do the epistemological views of David Hume and Immanuel Kant differ regarding what counts as “knowledge” in AI?

A) Hume believed in religious revelation, while Kant believed in mathematical probability.

Correct Incorrect Correct answer

B) Hume viewed knowledge as "correspondence" or patterns (A follows B), while Kant viewed knowledge as "causality" (the underlying reason why).

Correct Incorrect Correct answer

C) Kant believed only computers can have true knowledge, while Hume believed only humans are capable of it.

Correct Incorrect Correct answer

D) Hume argued for a "human-in-the-loop," while Kant argued for autonomous legal personhood for machines.

Correct Incorrect Correct answer

What is a significant drawback of the “Human-in-the-Loop” (Learned Intermediary) strategy for AI accountability?

A) It makes the AI model significantly less accurate by slowing down the processing speed.

Correct Incorrect Correct answer

B) It is illegal under current international data protection regulations.

Correct Incorrect Correct answer

C) It can defeat the purpose of automation, as constant human monitoring removes the benefit of using the AI in the first place.

Correct Incorrect Correct answer

D) Humans are biologically incapable of understanding any output from a modern neural network.

Correct Incorrect Correct answer

Emotion AI Ethics

James 28/05/2026

Quiz Summary

Information

Results

Results

Categories

1. Question

2. Question

3. Question

4. Question

5. Question

About Us

E-Learning

EU Collaborations

EA Consulting partners with 28DIGITAL on several EU funded and co-funded projects.

ICARUS AI