Can you think of examples of biases that exist in algorithms?
Mod 04 – Algorithms Need Managers – Discussion
This discussion draws from the article “Algorithms Need Managers, Too”by Luca, Kleinberg, and Mullainathan, from the “AI, Analytics, and the New Machine Age” book.
Please post your comments on at least three of the questions from the list below. Every student need not cover all the questions, but as a class I hope we cover all the questions.
1. The authors identify algorithms being “extremely literal” as a fundamental problems in how they “behave” differently from humans. Do you agree? is this a solvable problem?
2. The authors identify algorithms being a “black box” as a fundamental problems in how they “behave” differently from humans. Do you agree? is this a solvable problem?
3. Another claim is that “algorithms” tend to be myopic, that is, focus on the data-at-hand and short-term outcomes. Should this be avoided? If so, how?
4. Comment on the limitations of algorithms including the one’s identified in the article, adding any of your observations.
5. Recently emerging academic literature and business press articles are focusing on “bias” in algorithms. Can you think of examples of biases that exist in algorithms? What are their sources? How can they be avoided?
(I WILL ATTACH SCREENSHOTS OF THE ARTICLE TO USE IN ORDER TO ANSWER THESE QUESTIONS)