Case Study: Algorithms and Race

Loading...

From the course by University of Michigan

Data Science Ethics

53 ratings

University of Michigan

Data Science Ethics

53 ratings

What are the ethical considerations regarding the privacy and control of consumer information and big data, especially in the aftermath of recent large-scale data breaches? This course provides a framework to analyze these concerns as you examine the ethical and privacy implications of collecting and managing big data. Explore the broader impact of the data science field on modern society and the principles of fairness, accountability and transparency as you gain a deeper understanding of the importance of a shared set of ethical values. You will examine the need for voluntary disclosure when leveraging metadata to inform basic algorithms and/or complex artificial intelligence systems while also learning best practices for responsible data management, understanding the significance of the Fair Information Practices Principles Act and the laws concerning the "right to be forgotten." This course will help you answer questions such as who owns data, how do we value privacy, how to receive informed consent and what it means to be fair. Data scientists and anyone beginning to use or expand their use of data will benefit from this course. No particular previous knowledge needed.

From the lesson

Data Validity

Data validity is not a new concern. All too often, we see the inappropriate use of Data Science methods leading to erroneous conclusions. This module points out common errors, in language suited for a student with limited exposure to statistics. We'll focus on the notion of representative sample: opinionated customers, for example, are not necessarily representative of all customers.

Choice of Attributes and Measures6:12

Errors in Data Processing8:22

Errors in Model Design8:01

Managing Change5:18

Case Study: Three Blind Mice4:35

Case Study: Algorithms and Race3:44

Case Study: Algorithms in the Office3:18

Case Study: GermanWings Crash5:50

Case Study: Google Flu5:16

Meet the Instructors

H.V. Jagadish
Bernard A Galler Collegiate Professor
Electrical Engineering and Computer Science

Most people who write algorithms aren't racist, and don't want to write

algorithms that have a racist consequence or make racist decisions.

In fact, race may be farthest from their minds when they wrote the algorithm.

Nevertheless, the way the algorithms actually operate

could have significant racial implications.

For an example, let's watch this video clip of Professor Latanya Sweeney,

leading researcher on computer privacy

[MUSIC].

>> One day, I'm sitting in my office, and a reporter was there, and

he types in my name into Google because we were looking for an old paper of mine.

And up pops this add that had implied I had an arrest record.

Meanwhile, I show him where the result is we were looking for.

He says, no, no, tell me about this arrest?

[LAUGH] And I said, well, I've never been arrested!

And he says, then why does it say this?

And we go on and on, and we began searching more and more names.

And we started noticing this pattern.

And he jumps to the conclusion,

it's because you have one of those black names, he says.

And I said, that's ridiculous.

It's a computer, computers can't be partial.

And I would spend hours and hours trying to show him that he was wrong.

And, eventually, I do this cross-country,

120,000 ad research project, when I actually learned that

it was absolutely true that these ads showed up implying an arrest.

It had nothing to do with whether or not the company or

there was an arrest record under that person's name.

But if you had a name that was primarily assigned to black babies,

you were more likely to get an ad indicating arrest than if you had a name

more assigned to white babies.

>> So lets think a little bit about why what we saw may have happened.

When search engines choose what ads to show, they do so based on the ads

that are likely to be of interest, that the users are likely to respond to.

And if they have figured out that somebody who has a name of Latanya,

there's somebody who's looking for a search term of Latanya.

Is likely to be interested in services that might be provided to

criminal defendants, such as a bail bonds service or

something that you might want if you were arrested.

That's a reasonable ad to show.

That's a,ad that this person, who you don't know, is more likely to respond

to than thousands of other ads that you could possibly be showing.

This decision is purely a statistical frequency

type of decision that the algorithm has made, with no

thought whatsoever towards race or even racial implications of certain names

like Latania being more prevalent in the African American community.

Nevertheless, there can be real implications,

as we just saw in this video clip.

And understanding these real implications of supposedly

neutral algorithms that are operating in a completely data-driven manner.

Is something that anybody who writes algorithms or

practices data science needs to think about and own the consequences of.

Explore our Catalog

Join for free and get personalized recommendations, updates and offers.

Coursera provides universal access to the world’s best education, partnering with top universities and organizations to offer courses online.

© 2018 Coursera Inc. All rights reserved.

Download on the App Store

Get it on Google Play