The summer before my arrival at Purdue, I was asked to teach a version of Intermediate Ethics of Data Science, having never done it before. Despite having had some guidance from people who had taught the course previously, my own effort was, by all accounts, an abject failure—the worst I’ve experienced in my ten years in the classroom. Nothing went well. Jokes landed flat (or weren’t perceived as jokes). I did a terrible job managing expectations. I chose a suboptimal textbook. I failed to anticipate various ways in which computer science students differed from the student populations with which I had more experience. And the assignments did not get the students excited about the material we were covering. Slated to teach the class again this coming semester, I needed to do something different. But what?
About two years ago, I learned that Georgetown’s Jason Brennan received a Templeton Foundation Grant for his project, “Markets, Social Entrepreneurship, and Effective Altruism.” The grant was aimed, in part, at encouraging instructors to adopt an assignment Brennan calls the “Ethics Project,” with which he has had profound success. (You can read about it here). In its original form, the project directs students to “think of something good to do, and do it.” Instructors awarded funds from the grant commit to using the assignment in their classes.
Given the project-oriented mindset of my computer science students, it seemed like a natural fit. Rather than having students complete case analysis assignments as I previously had (assignments that were newly beset with generative AI-related challenges), I decided to give the Ethics Project a try. I applied for and received, funds to support 30 group projects in my 150-student class. As a result of this generous funding, each group this fall may claim up to $1,000 in funding to support its project.
Still, if this was going to work, I knew I’d have to adapt the original prompt to center ethical issues with data science. A generic “do good” prompt was insufficiently tailored to the course’s objectives. But I also wanted to preserve the project’s open-ended nature.
This is what I came up with:
Your final project for the semester will be to identify an opportunity to address an ethical issue involving data science or to think of something good to do by applying data scientific methods—and then act on it.
In groups of 5, students respond to this prompt throughout the semester and submit a written report detailing their work. This report is evaluated on four dimensions (which evaluation criteria I adapt from West Virginia University’s Dan Shahar:
- Conceptualization: How well have you explained why you chose your project and its relationship to our course theme, the ethics of data science? Did you demonstrate an appropriate level of ambition for a project of this kind?
- Planning and execution: How successful were you at formulating a game plan, anticipating obstacles, making adjustments as needed, and carrying your project through to completion? Did you manage to make a substantial positive impact through this project?
- Theoretical integration: Did you succeed in bringing your activities into conversation with ideas we’ve covered in this course? Do you present a cogent perspective on your project’s significance and merits in the context of the theoretical debates we’ve explored? Have you demonstrated mastery of the relevant material?
- Self-awareness and thoughtfulness: How well do you defend your chosen criteria for evaluating your project’s impact? Have you provided a thorough and honest assessment of your overall performance? Have you demonstrated you learned something from this project that goes beyond slogans and truisms?
To encourage students to think broadly about difference-making, the course begins with a unit that covers the effective altruism movement, the problem of world poverty, and the role that various organizations (for-profit businesses, non-profit organizations, trade associations, and governments) might play in making a positive difference in the world. In this unit, my hope is to encourage students’ creativity by making clear that their project can take the form of a for-profit business plan just as easily as it can more typically altruistic interventions (e.g., non-profit work, charity, or political efforts).
Once this point about different approaches has been driven home, we spend the rest of the course taking up topics like AI ethics and alignment; algorithmic bias and impartiality; data privacy and surveillance; social media addiction and well-being; and free speech, fake news, and content moderation. These units begin by pitching an application of data science as a solution to a social or moral problem—as many students will do in their final projects. They continue by exploring the new problem(s) to which the solution gives rise, unintended and sometimes unforeseen by the technology’s developer. The goal is to encourage students to think about the unintended consequences various real-world interventions have invited as they work on their own projects to make a positive difference.
The course culminates with a showcase of six student presentations—winners of the Purdue Data Ethics Prize. This is an opportunity for students to learn from and engage with successful projects by groups in other recitation sections and an incentive for students to take their project work seriously.
In addition to various components of the ethics project, student course performance is determined by three content-based quizzes, a short writing assignment exploring the promise and dangers of LLMs like ChatGPT (adapted from an assignment designed by WVU’s Brian Kogelmann), and topical engagement on an AI-moderated discussion forum.
I hope that this version of the class will help students conceive of ethics less as a set of constraints and prohibitions and more as an opportunity to think in a sophisticated way about the impact of their careers and projects. Whether this ambition is realized is, of course, a separate question that time will sort out before long.
The Syllabus Showcase of the APA Blog is designed to share insights into the syllabi of philosophy educators. We include syllabi in their original, unedited format that showcase a wide variety of philosophy classes. We would love for you to be a part of this project. Please contact Series Editors, Dr. Smrutipriya Pattnaik via smrutipriya23@gmail.com, Dr. Brynn Welch via bwelch@uab.edu, or Editor of the Teaching Beat, Alexis LaBar via labaralexis06@gmail.com with potential submissions.
JP Messina
JP Messina is an assistant professor in the Department of Philosophy. He offers courses in moral and political philosophy, the ethics of data science, and the history of practical philosophy. In addition to teaching responsibilities in the philosophy department, Messina teaches in the college’s Cornerstone Program. Before joining the faculty at Purdue, he held research positions at the University of New Orleans and Wellesley College. He received his Ph.D. from UC San Diego in 2018. His work has appeared in several scholarly venues, includingPhilosophers' Imprint, the Canadian Journal of Philosophy, the Journal of Applied Philosophy, Politics, Philosophy, and Economics, Kantian Review, and the British Journal for the History of Philosophy. His first book, Private Censorship, is forthcoming with Oxford University Press in 2023.
