Ethical Automation


Winter 2019

Time: Tuesdays 1:30 - 3:20 pm

Location: MOR 221

Chief explorer: Johan Michalove


Office Hours: Mondays 2:00 - 3:00 pm, CSE 220

Please do not hesitate to write to Johan about any accommodations or questions related to readings or course material. Additional meetings are available by appointment.


This course will investigate the role of values in how practitioners in artificial intelligence (AI), such as engineers and researchers, evaluate and construct systems which are automating tasks currently accomplished by humans. We will look at the effects of automation in thought work as well as in mechanized systems, and how practitioners in automation can perceive their own work in relation to broader social contexts. We will explore questions related to AI bias, identity, and the dissemination of information, among others. These questions will be motivated with hands-on work and discussion.

The course is targeted at motivated undergraduates in computer science who wish to delve deeper into the ethical questions surrounding automation. I expect that students will come into this course with varying familiarity with concepts in AI and its cross-disciplinary applications. Over the duration of the course, students will critically evaluate the day to day practices of fields within AI, as well as think about issues concerning the governance of the systems they have a hand in building.


This course will provide a space for students to analyze and articulate their views concerning ethical issues facing AI technologies. The point of this class is not for the dictation of what is ethical (or the contrary), but rather for students to hone their skills in critically exploring the development and implications of AI systems. At the end of this course, students will have gained a broader conception of dilemmas in current AI technologies and will have a stronger framework with which to develop their own ethical responsibilities.

Class Format

In each class we will investigate a topic which relates to the ethics, governance and design of AI systems. Each section will have a pre-assigned reading and supplementary content for students to explore. Discussions will be largely student driven, so I expect students to arrive prepared with discussion questions which interrogate the assigned readings. I encourage students draw upon their background and previous discussion sections to engage with the complexity of the subject at hand. Questions can include clarifications of particular portions of the text, relationships between the reading and a current event, or an interrogation into the values and underlying premises on which the reading was based.

We will begin each class by having a group of two or three students present on the assigned readings and the day’s subject. Each presentation should briefly review the paper and then engage with it through a set of initial discussion questions, tie-ins with other readings, and relate it to other outside research conducted by students. Please sign up here for a presentation.

Presentations will be followed by small group discussions, and finally we will turn to addressing the issues present in the topic through a large, class-wide discussion.


Each week, students are expected to read the assigned article(s) and submit through Canvas a brief response of three to five sentences in length, due the Monday at 11:59pm before class (except for our first meeting). These responses should demonstrate understanding of the reading and should entail of the following: raise or answer questions, connect the work to previous readings, critique the work, respond to a previous post, etc. The responses will be graded on a did/ did not do basis.

Students will also have the opportunity to go further with the ideas they’ve learned in class through a final project. Students can work in groups up to three people on a final project. Projects can include leading a discussion on an issue of interest, propose a solution to a problem posed in class (including code, if so desired), identify key stakeholders of a technology, analyze possible harmful implications of a current AI system, policy proposal or set of beliefs surrounding AI, or clarify in detail a key concept or definition discussed in class. I encourage you to meet with me to discuss your project proposal. There will also be opportunities to work on the final project in class.

Students should submit a project proposal by the fourth week of class describing: 1) members of the group, 2) the title of the project, 3) a project description no longer than one page, 4) two or more sources they plan to reference. This will be due 1/28 at 11:59pm.

The final projects will be due at 11:59 pm on 3/4 via email to Johan. The final section (3/12/19) will be devoted to presenting and discussing student final projects. Depending on the number of projects, we will also use our final exam period (3/22/19 - 2:30-4:20 pm) for final project presentations.


This seminar will be graded credit or no credit. In order to receive credit, students must submit 7/9 write-ups, present at one of the seminars, and complete the final project. That said, the point of this course is not a grade. Students should attend because of the readings and discussions—not in spite of them.

Class Materials

There is no required textbook for this class. All readings will be available on the home page of the course website--the majority of which will be either scans or online articles. Weekly readings will be announced on the Tuesday prior to when they’re due. I encourage students to share with me and the class readings they come upon.


1/8: Introduction



1/15: (Un) Ethical Machines



1/22: Reality Distortion I: Manifestos



1/29: Rationality, Operationalization and the Margins

2/5: Value Sensitive Design



2/12: Algorithmic Bias and Fairness



2/19: Engineering Culture



2/26: Reality Distortion II: DeepFakes and Democracy



3/5: Labor



3/12: Project Presentations

Check out: