CS 300-A - Human-Centered AI

Fall 1 (September–October) 2025

Course Description

Companies are investing billions of dollars in visions of what AI can do for people. Are these visions aligned with human flourishing? How can visions encounter reality? This project-based course teaches students to collaboratively discern the visions, assumptions, and design decisions behind current AI systems, while they design, implement and evaluate their own AI-powered applications with a human-centered approach. Students will practice creating systems that augment human capabilities rather than replacing them, with a focus on practical skills across the full development lifecycle and an emphasis on technology that promotes wholeness, well-being, and right relationships between people.

Student Learning Outcomes

By the end of this 7-week course, students will be able to:

  1. Analyze existing AI systems to identify underlying values, assumptions, and potential impacts on human flourishing
  2. Design AI-augmented user experiences that enhance rather than replace human capabilities and promote human well-being
  3. Implement functional prototypes integrating AI models with interactive interfaces
  4. Evaluate AI systems using both quantitative and qualitative user-centered methods
  5. Apply ethical frameworks such as the Design Norms and virtue ethics to AI development, considering privacy, transparency, user agency, and social impact
  6. Collaborate effectively on technical projects with diverse perspectives and skills
  7. Communicate design decisions through documentation and presentations

Prerequisites: Students should have taken at least one (ideally two) classes that involve either programming or enterprise software systems (e.g., INFO 301) – at Calvin or otherwise. Familiarity with web development concepts is helpful but not required; readings and resources will be provided to support learning.

Core Value: “Check against reality”

When generating is easy, checking is precious. We aim for ample opportunities to check our (AI-informed) assumptions against reality, so we can build systematic skills for learning from these encounters.

  • Industry: How do existing products connect human needs with AI capabilities? What was their PoC, how might it have failed, and how do they eval now?
  • Our Design Projects: We’ll rush through prototyping our ideas so that we can focus our effort on learning from how people interact with our technological ideas.

Grading

The assessments in this course have two goals: (1) to form a community where we can think deeply about human-centered AI and (2) to help each student form a portfolio of work that can outlast this class.

At the end of the course, you will write a brief self-assessment that includes a proposal for a final grade. You will discuss it with the instructor to at a final check-in meeting.

Components that will be considered in your final grade include:

  • Project (see below)
  • At-home practice activities
  • Thoughtful engagement in class activities
  • Engagement with readings and discussion forums
  • Giving and receiving feedback
  • Check-ins (on paper, every class meeting)

We will be using Perusall for readings for asking questions and discussing readings together. We use it because it can help build critical thinking skills, engage with peers, and build metacognition. We will use its automated grading system as a working example of a supposedly “human-centered” AI system and evaluate its claims.

AI Use Guidelines

We drafted these together in class:

  • Use AI together.
  • Don’t just submit AI’s output as your own
  • Document how you used AI
  • Push back on it. It’s probably not wrong but misguided.
  • Ethical considerations: use it for ideas but not judgments.

Course Structure

As a Fall 1 course, we meet a total of 19 times. This course will be very hands-on, with some activity most every day. Details are given on the meetings page.

In our readings, we will also engage various perspectives on the active debate about whether today’s AI is good, or good for us.

Primary Project

Students will conduct an integrated analysis of AI system design through three connected activities:

  1. Critical Analysis of some existing AI system: examine how the design choices that went into it impact human flourishing; envision alternative approaches
  2. Technical Evaluation of a simplified version (that you build) of the system’s core AI functionality
  3. User-Centered Design: prototype a complete interaction and iterate on it through feedback of various kinds.

Students may choose from provided template projects or propose their own with instructor approval. Final presentations (final 3 days of class) will be open to the broader campus community.

Details about the project are available on the projects page.

How does this differ from CS 375/376?

I also teach CS 375 and 376. Those courses emphasize how modern AI systems work, breaking it down into the math and programming details needed to interface with these systems at a low or medium level. The focus is on the AI, with some considerations of the human context.

This course inverts the emphasis: we focus on human contexts of AI, breaking down the product design of AI systems into what AI capabilities they depend on and how a company can learn whether it’s valuable for people.