DES 191B  ·  Spring 2026  ·  UC Davis

You never start
with AI.
You never end with it.

A ten-week studio where AI is a thinking partner, not a shortcut. The tool comes in the middle. What happens before and after is yours.

The course

A methodology
for this moment.

This course takes a position. AI is already part of how designers work. The question is not whether to use it, but how to use it without losing the thing that makes the work yours: the question you chose, the judgment you developed, the thing only you could have made.

The risk this course is designed against is a kind of funnel thinking: students converging too quickly on predictable, AI-shaped solutions, producing similar work from the same systems. The course does not counter this by restricting the tools. It counters it by insisting on the steps that surround them.

Over ten weeks, students build a methodology, not a toolbox. Week 1, they make something strange with AI and write about the first time they encountered it. Week 10, they return to that writing. The gap between those two moments is the course.

AI is never the first move. It is never the last. Everything between is where the work happens.

The framework

GENIE
a loop, not
a checklist.

By the midpoint of the quarter, students have already been running a five-phase process without knowing its name. Week 5 is when it gets named. GENIE is not new to you. It is the name for what you have already been doing.

GENIE is not a productivity tool. It intervenes in the design process itself. AI enters at Implement as a mid-process partner, calibrated differently in divergent phases than in convergent ones. What opens and closes every cycle is human judgment. Hover each phase to open it.

G
Gather
Research before making. Define the territory before picking a direction. Bring your own references, intuitions, and questions first. Then AI synthesizes sources, surfaces gaps, names what the work is already circling. Nothing is generated yet. The student leads; the tool organizes.
E
Explore
Students cannot move forward with a single solution. The goal shifts from finding the right output to understanding the space of possibilities. AI generates in parallel. HMW questions change what the problem is. What you curate and annotate shows you what you actually see.
N
Narrow
The convergent phase. Human judgment is the primary tool here. Name the specific person, the specific moment, the specific tension. Make a public commitment. The agent behavior shifts in convergent phases: it guides differently than it does when the work is still opening.
I
Implement
The inner loop: Generate, Experiment, Notice, Inspect, Export. The agent enters as a skeptical thinking partner, not a compliant one. It questions assumptions, isolates decisions, and makes the limitations of what was generated visible. You bring material. The tool probes it. You decide what to keep, discard, or transform.
E
Evaluate
Close the loop with reflection, not output. What did the tool get wrong, and what did that teach you? Where did imperfection become the most useful signal? What do you know now that you didn't in Week 1? The loop ends with the student. The tool has no voice in this phase.

The quarter

Gather · Explore · Narrow
Wk 1
Arrival
Who is here. What AI is and isn't. The question of authorship, before anyone makes anything.
AI Portrait. You By Me. Written reflection on first encounter.
Wk 2
Contact
First real encounters with generative tools. Two models compared. A before/now artifact: what did you think AI was, before this?
Before/Now artifact. 200-word written reflection. 2-minute video.
Wk 3
Pressure
Configuring a thinking partner using Six Levers. Research with AI. The counter-hypothesis: arguing against your own frame.
Thinking Partner configured. Six Levers: role, purpose, constraints, behavior, focus, confidence.
Wk 4
Locate
Tension statement. HMW questions. Intervention canvas. Midterm pitch: a public commitment to a direction everything else is measured against.
Midterm Pitch. Tension statement. HMW questions. Intervention canvas. 8–10 slides.
Implement
Wk 5
Engine Reveal
GENIE is named as the methodology students have already been running without knowing it. Thumbnails. The first implementation loop.
First full implementation loop. Generate, Experiment, Notice, Inspect, Export. Thumbnails.
Wk 6
Spine
Guest lecture: Felipe Luize (Principal Designer, YouTube). The WHY statement that organizes all four deliverables. Concept Brief written without AI first, then stress-tested with the thinking partner. Escalation ladder: Generate, Select, Refine, Control.
Concept Brief. Logline. Mechanism. Written alone first, stress-tested with thinking partner.
Wk 7
Production
Student case study presentations. Entry point sketch before any tool opens. Production workflow: Krea, Firefly, Photoshop. Hero image. Visual soul: three words, a not-brand, what it would say and what it would never say. Brand board.
Visual Soul. Brand board. Three words. A not-brand. Entry point sketch before any tool opens.
Wk 8
Form
Guest lecture: Rocket Drew (The Information). Visual seed: a generative paragraph that governs every image made after. Twelve shots tested across four categories of image-making.
Visual Seed paragraph. 12 generated images across four shot types.
Wk 9
Finish
Storyboard. Voiceover script. Video rough cut, then final cut: the WHY deliverable, told in four beats.
Final Video. Four-beat storyboard. Voiceover. Produced with narration and music.
Evaluate
Wk 10
Release
Narrative Figma page. Final presentation. The loop closes: what do you know now that you didn't in Week 1?
Narrative Figma page. Complete process documentation, made visible and accountable.

What I Know Now

Student reflections, Spring 2026
I started this course thinking AI would make things faster. It made things harder, in the right way. I couldn't hide behind a tool. The tool kept asking me what I actually meant.
Gabrielle Benaquista  ·  Overhand
The moment I realized the AI was wrong, and I could tell it was wrong. That was when I started trusting my own judgment again.
Aidan Fontano  ·  Identity Core
I thought I was making a project about memory. I was actually learning what it means to choose a version of something and own it.
Bella Chao  ·  What Remains
The best thing GENIE taught me: if you can't say what you're doing without the tool, you're not ready to open the tool yet.
Gabrielle Benaquista  ·  Overhand

Instructor

Jean H Menezes

Design · UC Davis

I built this course because I couldn't find a methodology for this moment: one that took authorship seriously, treated AI as a thinking partner, and didn't collapse into either panic or enthusiasm. The rule came first. You never start with AI, and you never end with it. Everything else followed from that.

GENIE was built inside the course, not before it. I named it in Week 5, after students had already been running the loop without knowing it. That sequence was intentional. The framework only lands when it names something the student already knows in their hands. It is not a productivity tool. It intervenes in the design process itself, which means most of its important effects are about how students learn to see, question, and revise.

The boundary I kept returning to in building this: AI can expand what is possible, but it should not replace the need for judgment. A responsible process is one where students use these systems to explore further, not to think less. The work here came from students who held to that distinction.