Using and Understanding AI

AI is changing how people write, study, create, and work. You can ignore it, panic about it, or learn how it actually works and decide for yourself what to do with it. This course is for the third option.

You'll spend the semester making things with AI and taking it apart to see how it works. You don't need any technical background to do either.

Registration requires approval. To apply, have a short conversation with an AI that gets to know you. We read every one and are putting together a class drawn from across the university.

Interested? Apply by Telling Us About Yourself

Applying is a short conversation with an AI about who you are, how you use AI now, and what you'd want to make. It's not a test, and there are no wrong answers. We read every one.

Start your application

Make Real Things

Most weeks, you build something. Last spring, students generated images, video, and full songs from a text prompt. They designed, built, and published their own websites in a single class meeting. They turned notes from their other courses into practice tests, then took a hard look at which AI study strategies actually help you learn. They handed a real dataset to an AI and kept asking questions until something interesting came out.

None of it required writing code. You describe what you want, look at what comes back, and push until it's right. The semester ends with a final project of your own choosing, built across two in-class workshops, plus a short video about it.

Understand What's Underneath

Making things is half the course. The other half is figuring out what is going on underneath.

You'll watch a language model predict the next word, turn up its temperature to see what changes, and compare it to a Markov babbler, a much simpler model that gets surprisingly far. You'll explore a single artificial neuron, watch a small network learn to read handwritten digits, and rate AI responses yourself to see how a raw model becomes a helpful assistant. An early scavenger hunt sends you after AI's failures: hallucinations, flattery, memory lapses, and overconfidence.

Then come the big questions. Can AI be creative? (We watch the AlphaGo documentary and argue about it.) Does it understand anything? What does every response cost in data, energy, human labor, and infrastructure? What does AI mean for work, relationships, safety, and human flourishing?

A Semester at a Glance

The course meets Tuesdays and Thursdays. Here is one week from the middle of the semester:

Week 5 · February 16–18, 2027

Tuesday · Lab

Study Guides

Bring materials from another course you're taking. You'll use AI to build practice tests and a study aid for it, then compare notes on which strategies actually help.

Thursday · Discussion

Does AI Understand?

Before class, read two articles, including Ted Chiang's "ChatGPT Is a Blurry JPEG of the Web," and talk them through with an AI that checks you're ready. In class, argue it out in small groups: do AI systems understand, or do they compress?

And here is how the whole semester unfolds:

1. Weeks 1–3

   Getting Started

   Hunt for AI's failures in a scavenger hunt, take your first conversational assessment, and make images, video, and music in the creative media lab.

2. Weeks 4–8

   How AI Works

   Next-word prediction, neurons and networks, embeddings, and the hardware and data behind it all, alternating with labs on studying and data analysis.

3. Weeks 9–12

   AI and Society

   Work, companions, safety, and where AI is heading. Learn how models are trained to be helpful, and build and publish your own website.

4. Weeks 13–15

   Final Projects

   Build a project of your own choosing across two workshops, step back to ask what AI means for human flourishing, and make a short video about what you made.

This follows how Spring 2026 ran, and topics and dates will change. The draft syllabus lists every meeting.

Who It's For

Anyone curious about AI. No programming or technical background is required, and no major is a better fit than another.

Skeptics are welcome. The syllabus says plainly that these tools have real costs and that reasonable people object to them. You will be expected to use AI throughout the semester, but the goal is for you to judge it from experience, not to sell it to you.

One practical note: CS 199 UAI is a pure elective. It does not satisfy any degree, program, or general education requirement.

How It Works

Class time is active. Instead of sitting through lectures, you work with a partner, discuss in small groups, and share out with the room.

Before each discussion, you read something short and talk it through in a preparation chat with an AI that checks you're ready to contribute. Instead of multiple-choice exams, you take conversational assessments, where an AI interviews you about what you've learned, the way a person would in an oral exam.

For Spring 2027, we may also add short weekly or biweekly quizzes on the readings, taken at the Computer-Based Testing Facility ⧉. The goal is to make sure everyone arrives ready for our in-class discussions.

Questions?

Ask below. The answers come from an AI that has read this page and the draft syllabus. This chat is not part of your application. For anything it can't answer, email Geoff at challen@illinois.edu.

Stay in Touch

Read the draft syllabus, or browse every meeting from Spring 2026 in the course archive.