- How would you describe the thesis (or theses) of this article?How does that relate to the title? And why
would we discuss this issue in an applied ethics class (1a-2a) - How does he restrict the scope of his claim? (1b) Is he right to do this in your view?
- The author proposes 5 possibilities for why we would want XAI. Of the five, which do you find the most
convincing? Which is the least convincing? (2a-4b) Would you add anything to the list? - What is the broadest distinction between methods of interpretability? (4b-5b) Put this distinction in your own
words. - What is the main analogy for “text explanations” (5b-6a)? Could this be successful, in your opinion? (7b)
- Why is it problematic to claim that, strictly speaking, linear models are more “interpretable” than neural
networks? (7a)