CSE 481L - Human-AI Interaction Capstone - Spring 2025

Course Overview

Course Description

This capstone explores the design and development of interactive technologies that integrate language models in providing AI-based interactive capabilities. Students will explore creating human-AI interactions that address real-world needs and tasks, applying principles from human-centered design in support of those interactions. Course activities will emphasize student-led discussions of current challenges and approaches, a group project exploring vertical prototyping of human-AI interactions, and examination of and reflection on designing with and for this emerging space.

This capstone focuses on human-centered design with and around AI:

  • We assume existing experience with human-centered design methods.
  • We assume existing experience with or the ability to learn frameworks for developing interactive applications.
  • We do not assume or expect existing experience developing with or around AI.
  • We do not assume or expect engagement with development of the underlying AI.

Given these expectations, the capstone publishes a pre-req of CSE 340, CSE 440, HCDE 318, or INFO 360.

Course Staff

James Fogarty

he/him

Mingyuan Zhong

he/him

Course Coordination

Course information and material will primarily be provided on this website.

Clarifications and updates may be issued as part of class. Although we will aim to ensure such updates are also captured on this website or in email, students will be considered responsible for announcements made in class.

Canvas and Drive will be used when appropriate for limiting access to course information and materials.

Assignments

Course assignments consist of:

  • Paper Presentations

    Student-led in-class presentation and discussion of research papers exploring human-AI interaction.

  • Technology Lab

    An self-defined initial technology lab, integrating an interface framework with AI-based capabilities.

  • Project

    A self-defined group project, exploring vertical prototyping of human-AI interactions in the context of real-world needs and tasks.

Grading

Strive to do good work because you care about your own opportunities to learn, including the opportunities this course provides working with a group in an intensive project.

Design is an inherently subjective practice, and so grading in this course is necessarily subjective. The stated project requirements are often the minimum, intended to leave room for groups to earn strong grades through strong work.

Assignment Grading

Individual assignments will typically be graded on a scale from ⭐⭐⭐⭐⭐ to ⭐.

This is intended to be roughly be interpreted as:

  • ⭐⭐⭐⭐⭐: Exceeds Expectations.

    Such work is very strong. This is above what is expected and will not be a common grade. This allows course staff to indicate appreciation of very strong work, going above already high course expectations.

  • ⭐⭐⭐⭐: Meets Expectations.

    Such work is strong. The course sets high expectations, and course staff feel this work meets those expectations. For example, the course staff is unlikely to have any specific objection to the work. The course staff may also not have any specific suggestions for improvement.

  • ⭐⭐⭐: Near Expectations.

    Such work is reasonable, but seems to leave room for improvement. There might be some clear concern or limitation, or there might be several issues which are individually minor but collectively a concern or limitation. Alternatively, the work might be appropriately executed, but may seem to have lacked ambition relative to expectations.

  • ⭐⭐: Problematic.

    Such work seems to fall short of the learning goals of the course. It may have met some minimal standard, but did not demonstrate the initiative and leadership expected in the course.

  • ⭐: Insufficient.

    Such work does not accomplish the intended goal of the assignment.

Assignment grading will focus on the work itself, not yet attempting to account for any group or individual adjustments (e.g., for late submissions, for contribution). Any such adjustments will be applied later as part of course grading.

Course Grading

Late Policy

We generally expect all assignments will be submitted according to assignment deadlines. If there is a constraint or an emergency impacting your ability to submit an assignment, contact the course staff before the assignment deadline by email to the course staff list.

We will do our best to be accommodating amidst requirements and expectations of the course (e.g., expecting clear and early communication regarding constraints, understanding that assignments support engagement throughout the course, understanding that missed assignment deadlines may undermine assignment utility or opportunities for feedback).

Regrade Policy

Any request for grade clarification or a re-grade must be submitted by email to the course staff list. Requests will only be considered within 7 days of the release of a grade.

Upon calculation of final grades, we will explicitly check for any particularly unfortunate effects of rounding (e.g., adjusting a pre-rounding final grade of 3.649 to instead give a 3.7). This is intended to provide peace of mind that final grades were not defined by any minor aspect of grading.

Expectation of In-Person Participation

The course heavily emphasizes required in-person activities, including student-led presentations and discussions. Remote participation will not be supported. Full engagement with and contribution in group project activities also requires in-person participation. In-person participation is therefore a required component of the course, and will be included in participation grading. In-person participation can also further impact grading of other course activities.

If you will be absent for any reason, contact the course staff before class by email to the course staff list. We will track absence-related communication as part of the expectation of in-person participation, including any context you want to share about the need for an absence. We will do our best to be accommodating amidst requirements and expectations of the course.

COVID-19 Safety, Recordings, and Zoom

There may be reasons you cannot or should not participate in-person, including COVID-19 exposure or symptoms. Our goal is for this course to provide an opportunity for learning without undermining safety, and we all share responsibility for preserving safety of this learning environment. Please review and adhere to COVID-19 related guidance, including:

https://www.ehs.washington.edu/covid-19-prevention-and-response/covid-19-illness-and-exposure-guidance

We will aim to record course sessions (e.g., unless a guest indicates they do not want to be recorded). Recorded sessions will be available for review within Canvas. However, we expect recordings will be of relatively poor quality and they are obviously an incomplete proxy for a discussion-focused course. Nevertheless, recording may sometimes be the best available option.

We may also use Zoom as an accommodation (e.g., due to illness, due to requirements of a guest). Zoom participation will use meeting information provided by email and in Canvas.

Accommodations and University Policies

Additional university policies apply (e.g., regarding Academic Integrity, Conduct, Disability Resources, Face Coverings, Religious Accommodations):

https://registrar.washington.edu/curriculum/syllabus-guidelines

Please do not hesitate to contact the course staff regarding these or any other accommodations. We are happy that you are here and want to support your engagement in this learning environment.

Contributing

This course website lives on GitHub:

https://github.com/uwcse481/web-cse481l-25sp