UNIV 3275 Spring 2020: The Ethics of Artificial Intelligence
Class meets T/Th, 2:35 pm – 3:50 pm, Rand 8
Douglas H. Fisher (Computer Science; Computer Engineering; Communication of Science & Technology) office hours T/Th 11:00 am – 11:30 am FGH 388
Haerin Shin (English; Cinema & Media Arts; Asian Studies), office hours T/Th 9 am – 10:50 am in Benson 412
This is a 3-credit course, and our expectation is that each student will spend approximately 9 high-quality hours on it per week, including 2.5 hours of in-class time, and therefore 6.5 hours outside of class.
The graded requirements for the course include weekly individual blog posts, individual blog responses, weekly group assignments, a late-term exam, and an end-of-term group project. A group poster is required.
The course requirements, their grading weights and descriptions, are given below.
Course Requirements and Grading
- Attendance and active participation in discussions (10%)
- Weekly group assignments (25%)
- Virtual Forum posts and responses (20%)
- Late-term Exam (15%)
- Immersive Group Project (30%)
Readings, film, and materials in other media will be required of students before the class in which the material is discussed (see Schedule). Understanding of these materials will also be reflected in weekly group assignments and discussion forum posts/responses, as well as a late-term exam and group project.
Attendance and active participation in discussions
A very important way of participating in the course is to attend class and participate in discussion. Class attendance is easily measured, and participation in discussion is more subjective (and can include participating in other active ways, like live tweeting on the discussion rather than talking out loud). We won’t quantify the quality of “active participation” here, but the quality of discussion will be informed by demonstrated knowledge of the week’s assigned and student-identified material.
Regarding attendance, which is crucial in a class such as ours, a student will be allowed up to three unexcused absences, with each absence thereafter counting 0.6%. We will take attendance using Top Hat. If a student is traveling or unwell, we can set up a connection using Zoom that will allow them to “virtually attend”. We are pretty much softies if you contact us ahead of the class you are going to miss, with an understandable reason for absence.
We will excuse classes missed before a student adds the course, and consider outstanding participation in discussion as potential compensation for a past missed class.
Time required for class attendance each week is 2.5 hours.
Time required for assimilation of assigned material before classes is estimated at 2.5 hours.
Over the semester there will be approximately 8 group assignments, 4 veering more toward the technical and 4 that fall more on the literary side. The group size, which will be 3, is designed to ensure that each student will have an opportunity to interact with as many other students as possible. Group composition will change from week to week.
There is an important additional cost to missing class, particularly Thursdays when Group Assignment teams are formed. We will not facilitate your addition to a team if you aren’t there or lack a good reason for missing out. If you do not take part in a group, then you cannot get a Group Assignment grade (not even if you submit the assignment individually, unless under exceptional circumstances).
We estimate that the time required for the group exercise is 2 hours, with 1 to 1.5 hours required for the interaction with technology and literature, and 0.5 to 1.0 hours required for the short-answer questions.
Group Technical Assignments
Technical group assignments are intended to explore “under-the-hood” aspects of AI, as well as exploring the ethics of an AI application. This class does not require a technical background and is open to all majors; the technical assignments are designed accordingly. Examples of technical assignments may include running an off-the-shelf, easy-to-use machine learning system on judicial decision-making data, or running an easy-to-use intelligent vehicle simulation that illustrates different “moral” choices in accident scenarios.
Following the technical exercise, students will collectively fill out answers to questions about the exercise that is intended to seed deeper reflection.
Group Literary Assignments
These assignments are intended to reflect on “preempting the future” by considering the future of AI in a variety of situations found in literature, and to imagine how this future might be shaped. After reading an excerpt from a literary piece students will collectively answer questions, or answer in other formats, such as writing a short script about a possible future.
Virtual Forum posts and responses (See below for format as of Week 5) Each week, and for each reading from that week, students will make an individual post that proposes a possible exam question and a short answer to that question based on the reading and their own reflection. Each week students will respond to at least two such questions, either with an alternative answer to a proposed question or a modestly revised question. Proposed questions can be no more than 30 words, and proposed answers can be no more than 50 words, and no less than 20 words. 80% of the late-term exam will be taken from these posts and responses. We estimate that the time for posts and responses to be 1.0 hour per week for writing and reflection, with the understanding that reflection time is built into all components of the course.
Virtual Forum posts
Each week, following the group technical exercise, students will post an individual reflection of approximately 250 words on the week’s material to date, to include the group technical exercise. By “approximately 250 words” we mean to suggest a guideline no less than 200 words and no more than 300 words. The virtual forum post will cite sources and acknowledge influences in the post.
We estimate that the time for posts to be 1.5 hours per week for writing and reflection, with the understanding that reflection time is built into all components of the course.
Students are required to read other student posts and to respond to one or more of those posts in a substantive way. There is no word limit on responses. The response might be a long analysis or a short description of a relevant source, which the responder has read or watched or otherwise interacted with, with appropriate citation and URL, if available. There are different forms of high-quality responses. A response can be a response to the original post or to another response.
We estimate that the time for responses to be 1.0 hour per week for writing and reflection.
Typical Week (Week k)
- Finish studying materials for week k that were assigned in week k-1 (i.e., readings, film, or in other media) by Monday of week k;
- Forum posts are due Mondays of week k at 12:00 pm (Noon)
- Small group discussion and report back in Tuesday class of week k
- Weekly Group Assignments for week k are due Wednesdays of week k at 12:00 pm (Noon)
- Form groups for group assignments of week k+1 in Thursday class of week k
- Guest (faculty) discussant in Thursday class of week k
- Start studying materials for week k+1 (if you haven’t already) and planning/doing group assignment meet ups for week k+1
- Forum post follow-ups are due Saturday of week k at 12:00 pm (Noon)
Because all materials are posted for each week at the start of the semester, please organize your study schedule to ensure timely observance of weekly deadlines. This may require, for example, not waiting until after Thursday class to start on the following week’s assigned materials.
Late-term Synthesis Exam (Essay Format)
An exam will be given in the second-to-last week of the semester. 80% of the points total will be taken from the questions class members have proposed on the Virtual Forums.
A late-term take-home synthesis exam will include questions in a variety of formats, to potentially include short answer, longer written response, multiple-select (which will nonetheless require non-trivial reasoning before selecting options), and essay. The intent is that the exam will invite students to synthesize across the material.
Final Group Project
The final group project will be of each group’s own design, and including a collaboratively-written report (e.g., that describes any computational AI studies, related work, motivation, and the like) and a collaboratively-prepared poster that will be used in a class conference. Teams will be formed to ensure interdisciplinary to the extent possible. The content of the project will vary dependent on student interests; it could be on areas precisely covered in class (e.g. a survey and analysis with therapy robots) or on topics not substantially covered in, but inspired by and related to class content under the same broad umbrella of AI sub-areas (e.g. humanoid household-work robots).
Consultations between student groups and the instructors will accompany group design, development, and conclusion of the project.
Other typical course responsibilities, described above, will be adjusted to accommodate student needs to work on the project in the last quarter of the semester.
While the final group project is collaborative, an individual student reflection of the project is also required. More on the final group project is found here.