Skip to content

ECSE681: Colloquium on Electrical Engineering on Responsible AI

General Information

Offered at McGill University

Semester: Winter 2025

Instructor: AJung Moon

Office Hours: by appointment only

Course Credit: 4 credits

Expected course work outside of the lecture hours: 12 hours/week 

Prerequisite: None

Lecture Format: In person by default with a few exceptions (e.g., remote talks/guest lectures)

Lecture Time: 9:35 am – 11:25 am on Wednesdays and Fridays

First Lecture: Wednesday, January 8, 2025

Physical Room: Wong Building 1030

Tutorial: None

NOTE: In the event of extraordinary circumstances beyond the University’s control, the content and/or evaluation scheme in this course is subject to change.

About the course

With the rapid deployment of AI systems everywhere, the need to practice responsible AI has never been more important. To protect the public from possible harms stemming from AI systems, nascent AI regulation activities in the western society also demand AI systems to be audited, not only to ensure the system’s performance but also for its responsible design and deployment. But what does it mean to practice responsible AI? What do the terms AI safety, AI ethics, AI for Good and responsible AI all mean?

This colloquium is designed for students wishing to become the next generation of responsible AI practitioner. We will explore the rapidly changing domain together. Along the way, we will learn how to be a more positive part of the AI ecosystem in our own ways. In particular, we will explore this quest from the perspective of responsible AI practitioners, AI ethicists, or ethics-aware ML practitioner. We will engage in real world problems. read latest articles and classical texts.

The course is designed to be engaging and interactive. It includes seminar talks given by guest speakers, lectures by the instructor, and student-led activities. By the end of the course, students will be expected to have experiential, rather than an arms-chair understanding of AI ethics issues and assessment processes, and have a working knowledge of the role that responsible AI practitioners play in the AI industry.

Learning outcomes

  • A practical understanding of the role AI ethicists play in the industry
  • Ability to identify and analyze ethics issues specific to an AI system using existing tools/methods
  • Ability to communicate results of an AI ethics assessment clearly and effectively to multistakeholder groups
  • Ability to ideate sociotechnical solutions to AI ethics issues given the rapid pace of the domain

Topics list

  • AI ethics assessments and sociotechnical harms & risks
  • Facilitating conversations about values and ethics
  • On-going & new AI governance activities
  • Human-AI interaction problems
  • Responsible AI practice and profession

What to expect

Lectures

This is a highly interdisciplinary course covering highly popular, and dynamic topics. Many of the topics we will cover will be on contentious subjects that don’t have closed-ended “right” answers. As such, students can expect the lectures to include active discussions where students are expected to participate, think aloud, and share opinions. To enable rich discussions with the AI community within and outside McGill, some of the lectures may coincide with seminar talks/reading discussions open to local AI stakeholders from outside the class. Some guest lectures may also be streamed online. Students registered for the course are expected to attend in person to all scheduled lecture slots unless otherwise stated.

Student Work

As a 4-credit course, students should expect to set aside an average of 12 hours of work per week toward the course work outside of the scheduled lectures. This is a rough rule of thumb, and will depend on your individual ability to read, explore materials independently and so on. Students are expected to come to each lecture having read any assigned readings or preparation before the lecture.

Course Project

As part of the course, you will have the opportunity to conduct and ethics audit of an AI system of your choice. The instructor will guide you through the process of AI system selection and report delivery. Students will be expected to deliver a report and a presentation from the audit as third party auditor of an AI company. Although not mandatory, students will be encouraged to build their public, professional portfolio stemming from their work on the course material and the project.

How we will communicate

This website is built to enable ease of course material from the instructor to a wider audience. All announcements for registered students and assignment material will be posted on myCourses. Students registered for the course should expect to submit all assignments via myCourses. Students also have the option to publish their project-related contents on platforms/format of their choice. The only requirement for students wishing to publicise their work along the way is to make the material is accessible to the instructor at the time of assessment and can be stored in a static/permanent format for possible course audit and related purposes.

Video Recording of Course Content

The default format of the course this semester is in-person. These in-person classes will not normally be recorded even if some guest lectures are streamed live. Video or audio recordings of the lecture by students for their individual or public use is not permitted. This is to establish a level of psychological safety needed for the students to freely discuss difficult and often emotionally charged topics in class.

Should the course need to be delivered online or in a hybrid format for any reason, much of the lectures and tutorials will be recorded and the recordings made available on myCourses afterwards. You will be notified through a “pop-up” box in Zoom if part of a class is being recorded. By remaining in sessions that are recorded, you agree to the recording, and you understand that your image, voice, and name may be disclosed to classmates. You also understand that recordings will be made available in myCourses to students registered in the course. During breakout sessions in class, only the breakout rooms where the lecturer visits may be recorded. If you don’t feel comfortable being recorded as part of the online lecture recording/tutorial, you can choose not to take part by logging off Zoom and watching recorded course components later.

Course text and references

As a seminar course meant to encourage exploration of responsible AI and related topics across disciplines, this course does not have a designated textbook. Rather, there are reading assignments related to different discussion items we will delve into as outlined in Schedule.

Method of Evaluation

The student evaluation for this course is designed to mimic the process of self-learning of interdisciplinary contents, facilitation, presentation, and reporting expected of responsible AI practitioners today. Students will have the opportunity to take the role of participant/audience/auditor throughout the course and encouraged to give feedback to their peers.

  • In-class activities/presentations (40%)

  • Short paper (10%)

  • Mini risk-assessment (10%)

  • Final project submission (40%)

Course Schedule

See Schedule