Group Project

AI 380: Human-AI Interaction · Fall 2026

Note

The group project is the core learning experience of AI 380 — worth 60% of your final grade. It runs in two sequential projects that take your team from initial user research through a fully deployed, user-tested conversational AI system.

Overview

This course centers on two sequential group projects designed to progressively build your human-centered design and conversational AI skills:

Project	Focus	Milestones	Weight
Project 1	Conversation AI Learning Environment (VR)	GP0, GP1	20%
Project 2	Conversational AI Assistants for Older Adults	GP2, GP3, GP4	40%

Teams of 2–3 students are assigned by Sep 5 (Last Day Add/Drop). Each team must submit a signed Team Working Agreement (GP0) before work on Project 1 begins.

Warning

Students who do not actively contribute to group activities will lose the grade for that activity. The instructor reserves the right to assign individual grades that differ from the group grade based on documented contribution.

Team Formation

Group size: 2–3 students per team
Formation deadline: Sep 5 (Last Day Add/Drop)
Interdisciplinary teams are encouraged — diverse perspectives produce more inclusive designs
Students unable to find partners should notify the instructor immediately

Team Work Agreement

Each team submits a signed Work Agreement (template on Canvas) as part of GP0, covering:

Team member names and contact info
Meeting schedule and communication norms (Discord, email, etc.)
Decision-making process
Conflict resolution strategy
Expectations for equal contribution

Project 1: Conversation AI Learning Environment

Overview

In this first project, teams co-design and build an interactive VR learning experience for older adults (ages 70+) using the course’s custom web-authoring tool. The goal is to support cognitive health, memory, or daily routine management through a carefully designed 9-screen VR activity validated on the Meta Quest 3.

This project develops skills in: needfinding, paper prototyping, digital authoring, VR interaction design, and older adult UX considerations.

Milestones

Milestone	Description	Due	Weight
GP0	Team Formation, Charter & Project Journal Setup	Sep 25	5%
GP1	Co-Designing & Authoring VR Interventions	Oct 2	15%

Design Considerations for Older Adults

All VR activities must incorporate UX accommodations for users 70+:

Large UI targets — buttons and hotspots with generous touch/pointer areas
High-contrast visual cues — colors and borders that remain legible with reduced vision
Gentle pacing — no time pressure; patient transitions between screens
Audio support — important instructions read aloud, not text-only
Low-stress error recovery — positive feedback regardless of initial response; no penalties
Minimal cognitive load — linear layouts, one action per screen

Authoring Platform

Teams use the course web-authoring tool to build and export VR lessons:
http://147.182.173.208:5173/

The platform allows teams to author 9-screen interactive sequences and export them for deployment on the Meta Quest 3. Instructor will provide onboarding in Week 9 lab sessions.

Example Activity Scenarios

💊

Medication Management

Help older adults learn to organize, sort, and manage daily medications — sorting pills by meal timing, understanding label instructions, setting reminder alarms.

🧠

Memory Exercises

Sequence-based activities for short-term memory practice — ordering morning routine steps, matching terms, or completing daily checklists.

🏠

Daily Living Skills

Practice navigating everyday tasks — preparing an emergency kit, planning a grocery list, or reviewing home safety procedures.

Project 2: Conversational AI Assistants for Older Adults

Overview

In this main project, teams ideate, design, build, deploy, and evaluate a conversational AI assistant aimed at addressing the specific needs of older adults (70+ years). Guided by the Design Thinking framework, teams work closely with older adults throughout the semester to develop solutions focused on one or more of these core domains:

Cognitive enhancement — memory support, medication reminders, cognitive exercises
Physical mobility support — navigation assistance, exercise guidance, fall-risk awareness
Social well-being — companionship interfaces, family communication tools, social activity suggestions

Milestones

Milestone	Description	Due	Weight
GP2	User Needfinding, Personas & Paper Workflow	Oct 16	15%
GP3	System Implementation & Containerization	Nov 20	15%
GP4	Public Deployment, Usability Testing & Final Docs	Dec 3	15%

Project Timeline

GP0

Team Formation, Charter & Project Journal Setup

Due: Sep 25 · 5%

Your team is formed, your working agreement is signed, and your public project journal website is live. This foundation enables all future project work.

Signed Team Working Agreement (all members)
Team project journal website (Home, Process, Deliverables)
Team members, roles, and project vision posted publicly

GP1

Co-Designing & Authoring VR Interventions for Older Adults

Due: Oct 2 · 15%

Your team has engaged an older adult (70+) in needfinding, produced a paper prototype of a 9-screen learning activity, authored it using the course VR authoring tool, and validated it on the Meta Quest 3.

Needfinding artifacts (interview/observation notes, photos)
Paper prototype (state diagram or storyboard)
9-screen VR activity authored on the course platform
Narrated Meta Quest 3 demonstration video
Authoring system export file or link

GP2

User Needfinding, Personas & Paper Conversational Workflow

Due: Oct 16 · 15%

Your team has conducted needfinding with older adult participants, defined a clear problem, built representative user personas, and mapped the full conversational workflow for your AI assistant on paper.

Needfinding summary (2+ older adult participants)
Problem statement and scoping document
2 user personas (grounded in research evidence)
Paper-based dialogue flowchart (multi-turn paths, fallbacks, human-in-the-loop)
Project website updated with all GP2 artifacts

GP3

System Implementation, Integration & Containerization

Due: Nov 20 · 15%

Your team has built a functional conversational AI assistant with Rasa CALM, wrapped it in an accessible web/mobile interface, and containerized the full stack with Docker.

Functional Rasa CALM agent (multi-turn dialogue)
Accessible web/mobile frontend interface
Docker container (Dockerfile + docker-compose)
Technical README with setup instructions
Project website updated with architecture and screenshots

GP4

Public Deployment, Usability Testing & Final Documentation

Due: Dec 3 · 15% · Presentation Dec 2

Your system is publicly deployed. You have conducted field testing with real older adult users and documented your full design process in a scholarly report and public presentation.

Publicly accessible live URL for the deployed assistant
Usability testing report (3+ older adult participants)
Academic poster and technical documentation report
12-min final presentation + 5-min Q&A + live demo
Individual contribution statement (submitted privately)

Design Process Requirements

Each milestone builds on the previous, following the double-diamond design process:

Discover → Define → Develop → Deliver
  GP0/GP1  →  GP2   →   GP3  →   GP4

Research Phase (Weeks 1–8)

Engage older adult participants (70+) for needfinding — not just assumptions
Conduct participant observation or semi-structured interviews
Use informed consent (template on Canvas)
Document findings with anonymized quotes and behavioral observations

Design Phase (Weeks 4–8)

Paper prototype conversational flows before building
Use Workshop 1 (Week 4) and Workshop 2 (Week 8) for structured ideation
Produce paper dialogue flowcharts for all major conversational paths

Build Phase (Weeks 9–11)

Use Rasa CALM as the conversational AI framework
Build in accessibility from the start (don’t retrofit)
Participate in Lab 2, Lab 3, and Lab 4 sessions for hands-on implementation support
Complete Docker containerization before GP3 deadline

Evaluation Phase (Weeks 12–14)

Conduct heuristic evaluation in Week 12 workshop with a peer team
Run user studies with at least 3 older adult participants (Nov–Dec)
Use think-aloud protocol + post-session questionnaire
CITI research ethics context applies — treat participants with care and respect

Contribution Statement Requirements

Mandatory for GP1 through GP4. See Assignments → Contribution Statements for the full requirement. Both a team-level breakdown AND a private individual statement are required for every group submission.

Peer Evaluation

At the end of the semester, all team members complete a confidential peer contribution survey on Canvas evaluating:

Teammate contributions to research, design, and implementation
Communication and reliability
Overall collaboration quality

Peer evaluation results inform individual grade adjustments within teams.

Project Resources

Resource	Purpose	Link
VR Authoring Tool	Build 9-screen VR lessons (GP1)	http://147.182.173.208:5173/
Rasa Docs	CALM framework reference	rasa.com/docs/rasa-pro/calm
Docker Docs	Containerization guide	docs.docker.com
Google PAIR Guidebook	AI product design patterns	pair.withgoogle.com/guidebook
Microsoft HAX Toolkit	Human-AI experience design	microsoft.com/en-us/haxtoolkit
WCAG 2.1	Accessibility guidelines	w3.org/WAI/standards-guidelines/wcag
Canvas	Submissions and grades	willamette.instructure.com/courses/8131

Project FAQ

Do we choose our own project topic within Project 2?
Yes. Teams choose the specific domain (cognitive, mobility, or social) and the exact problem to solve. The constraint is that the target population is older adults (70+) and the system must be a conversational AI interface.

Can we build a real system, not just a prototype?
Yes — and you are encouraged to. Using Rasa CALM to build a functional conversational agent is required for GP3. Technical sophistication is valued, but the primary focus for grading is the design process and evaluation rigor, not how many features you ship.

What if we cannot recruit older adult participants?
Contact the instructor immediately. The class Discord has a #participant-recruitment channel. The instructor can provide access to community partner networks. Do not simply substitute CS students for your target population.

How should we handle a teammate who is not contributing?
First, address it within your team using the Work Agreement. If unresolved, contact the instructor with documentation. Individual grades may be adjusted based on contribution evidence.

What technology is required for GP3?
Rasa CALM (or equivalent enterprise conversational AI framework), Docker for containerization, and a web or mobile frontend. Specific framework choices should be discussed with the instructor by GP2.

Can we use ChatGPT or other LLMs in our system?
Yes. AI-augmented prototypes are encouraged. Document your use of AI tools in your report, including what they were used for and any limitations observed.