COMP 250: UNDERSTANDING THE MACHINES: AI
Spring 2027 · Section 001 · CRN 26943
Generative AI has, in only a few years, become something every educated person must reckon with. Yet most ways of learning about it fall into two camps: popular accounts that explain what AI is without ever saying how it works, and technical courses that teach engineers to build it behind a wall of mathematics. This course occupies the missing middle. It carries students with no background in mathematics or programming from the first idea of a thinking machine all the way to the generative systems now reshaping daily life—rigorously and honestly, and without ever asking anyone to solve an equation. Wherever an idea would ordinarily be carried by an equation, we carry it instead with an analogy, a picture, or a worked example. The course is for everyone, and the word is meant literally: the music major and the historian belong in the room as fully as the future engineer or physician. Understanding these systems is no longer optional for an educated person in any field, and the most consequential questions AI raises are not, in the end, engineering questions at all. The course moves in six parts, each resting on the one before: from what intelligence in a machine could even mean (Part I), through how classical machine learning finds patterns (Part II) and how neural networks learn in layers (Part III), to the generative era of attention and large models (Part IV), and on to the data, society, and responsibility these systems force upon us (Part V) and what it all means for your own field and for thinking well alongside such tools (Part VI). Running beneath the whole course is a single lens—a spine of six questions to bring to any AI system: What problem is it solving, and who chose that framing? What data did it learn from, and at what cost? What kind of model is it? How was it trained? How is it evaluated? And what is its impact? The aim is not to make you an engineer; you will not leave able to build these systems. It is to make you fluent—able to meet any AI system, now or in the future, and to understand it, question it, and use it well, rather than fear it or trust it blindly.
- Credits
- 3
- Instructors
- Luay Nakhleh