Teaching Human-AI-Interaction

Course Structure

Human-AI Interaction (HAII) is taught over 12 weeks at Williams College to only undergraduates. The course is organized around 12 modules (i.e., topics), and each module includes: 2x pre-recorded lecture videos, 2-3 readings (1 research paper + 2 popular media), a ~7 question quiz on that module's materials, an 80 minute in-person class meeting called "Conference Section" for small group activities introduced by the lecture videos, and an 80 minute in-person class meeting called "Discussion Section" for student-lead discussions of the readings. The course is implemented in Canvas, referred to as "GLOW" at Williams, but I have made available some versions of the materials via Google documents on this website (apologies for any formatting issues!). Context for each course component is below, but details can be found on the Schedule and Syllabus pages.

• Lectures: Pre-recorded lectures for a module are posted once per week (two playlists, both posted Thursdays), which were recorded in < 15 minute sections and linked together in a YouTube playlist for no more than 50 minutes total, although usually around 30-40 minutes. Lectures would include thinking activities for the student to do prior to meeting in Conference Section. "Guest Lectures" (i.e., videos, or several videos from others on YouTube) were occasionally used to provide external expert insight on topics, but also to emphasize the relevance of the course content.

• Readings: Each module had either 2 research papers, or 1 research paper + 2 popular media readings, with the latter option being preferred by the audience of undergraduate Computer Science majors.

• Comprehension Quizzes: Prior to meeting in Conference Section, students completed a short 7-9 question quiz testing their understanding of the readings and lectures. It's important to focus on higher level concepts rather than difficult to remember details.

• Conference Section: Conference Section time is an 80 minute class period focused on: reviewing troublesome Comprehension Quiz questions, Activity Sessions with ~2 peers reviewing the activities from lecture, new Activity Session activities complementing the content, and returning to discuss as a group. Activities were generally organized around a set of editable Google Slides (many adapted from Training for Change), so students had a shared space to share their small group insights.

• Discussion Section: Discussion Section is an 80 minute class period in which the professor introduced tech topics that were happening right now, ~45 minutes of student-lead discussion of that week's readings, and ~15 minutes of student help hours at the end of class.

• Assignments: There are 4 assignments for this class and a final project (with 4 options from which to choose). They generally occurred every other week, although Assignments 1 & 2 were a bit lengthier. Pass/Fail Check-ins were implemented at the midway points to help students who tend to leave these assignments to the last possible second.

Course Catalog Description [link]

Artificial intelligence (AI) is already transforming society and every industry today. In order to ensure that AI serves the collective needs of humanity, we as computer scientists must guide AI so that it has a positive impact on the human experience. This course is an introduction to harnessing the power of AI so that it benefits people and communities. We will cover a number of general topics such as: agency and initiative, AI and ethics, bias and transparency, confidence and errors, human augmentation and amplification, trust and explainability, and mixed-initiative systems. We explore these topics via readings and projects across the AI spectrum, including: dialog and speech-controlled systems, computer vision, data science, recommender systems, text summarization, and UI personalization, among others.

Removed: Students will complete individual bi-weekly mini-projects in which they will design and build AI systems and components across a wide variety of domains. Students should be comfortable with programming; assignments will be primarily in Python. Prior experience with AI/machine learning will be useful but is not required. Students will also be responsible for weekly readings, pre-recorded lectures twice per week, short quizzes on the reading & lecturing materials, weekly synchronous [remote] small group classes, weekly discussion posts, and occasional presentations to the class.