PIXELATED: A Designer’s Guide to GenAI

BRIEF

Co-designed with Generative AI, PIXELATED serves as a roadmap for designers venturing into the field of GenAI for the first time. Through interviews with 20 fellow designers and secondary research, this issue delves into the ethical considerations, creative potential, and the delicate balance between human intuition and machine intelligence.

DURATION

10 weeks

SCOPE

Research
Publication Design
Bookbinding
Exhibition Design
Screenprinting

TEAM

ChatGPT Plus
Google Gemini
Midjourney
Dovetail AI

ADVISED BY

Chad P. Hall

PREVIEW

Publication + Screenprinting
+ Interaction Design

Spring quarter of my junior year, I had the privilege of joining MDes students in a graduate design studio to explore the rapidly evolving field of Generative AI. In this small, 6-person studio, each of us conceived an original project from the ground up, using GenAI tools as our collaborative co-designer(s).

My project, PIXELATED, is a three-part exhibition that captures the complexity and ambiguity of Generative AI within the design landscape. It includes:

A magazine publication, featuring secondary research and 20 in-depth interviews with designers + GenAI practitioners

A screenprinted poster

A theromochromic sheet, a metaphorical piece that uses heat-sensitive black ink to reveal various opinions from designers about GenAI

BACKGROUND

What is GenAI?

To work with GenAI as my co-designer, I needed to know more about its capabilities. I conducted thorough research on large language models (LLMs), neural networks, and the algorithms driving the generative process.

Landscaping

I was struck by how powerful GenAI has become in such a short time. Curious, and a little overwhelmed, I began to wonder if human designers and GenAI might co-work by combining our strengths — GenAI enhancing efficiency in the design process, while humans retain the essential empathy, research, and creativity.

So during the landscape analysis, I found myself naturally drawn to exploring GenAI tools that could support the design and prototyping stages.

From there, I wanted to get an overall sense of their strengths and weaknesses. I intentionally selected 3 GenAI tools from companies of varied sizes, conducted multiple rounds of testing, and documented my findings.
‍

Personal Project (Figma Plugin)

Wireframe Designer

Startup

Archie AI

Big Tech

Sora

The more I learn, the less I know!

As GenAI tools continue to develop and evolve, processing information at double or even triple the current speed, designers have very little time to adapt to this change. How can we keep up? Should we simply jump on the bandwagon? I realized the importance of addressing the role of GenAI in design and helping designers navigate this landscape, so we can understand what we can control, amidst all that is beyond our control.

PROJECT PROPOSAL

A Zine

I was drawn to the idea of making a zine because it’s a well-designed and a widely accessible format for sharing information. I wanted to achieve 2 goals:
‍

Interview designers working across various fields, and highlight a broad spectrum of thoughts and perspectives on GenAI

Conduct secondary research to guide the content narrative, addressing key topics from the interviews with a more research-informed, evidence-based approach

RESEARCH

Survey

I started with a general survey to collect early-stage data on designers' overall sentiment toward GenAI. To reach designers specifically, I mainly distributed the survey through personal connections, LinkedIn, and friends of friends within the design community. I received 27 responses in total.

Interviews
+ Thematic Analysis

Based on the survey responses, I conducted 20 interviews with participants, along with an additional interview with my professor, who works on GenAI and machine learning at Facebook Reality Labs. The interviewees represent a diverse range of design fields, including UI/UX, visual communication design, industrial design, and design education.

Next, I uploaded the interview transcripts to Dovetail and coded the data to identify themes that would shape the zine's narrative structure.

DESIGN

Pixelated Moodboard

When putting together the moodboard for zine’s visual design, the word “pixelated” naturally came to mind. The pixels carry different meanings depending on the context — they may refer to the digital pixels that designers move around in Illustrator, or the pixels that become more defined during the GenAI image generation process.

“Pixelated” not only describes the common ground between human designers and GenAI, but also alludes to the ambiguity of how both fields evolve. This concept solidified my visual language and clarified the message I wanted to communicate through the zine.

Collaborating with GenAI

Co-working with GenAI was a new experience for me, but very rewarding and fascinating. Through many trial and errors, I found a good balance in dividing up the work, combining our strengths and weaknesses to achieve the best result.

I broke the process down into 2 main stages:

01 Data + Structure

ChatGPT and Dovetail AI summarized my research notes into digestible bullet points. I went back in to review the data and refine the AI-generated summaries, and then organized them into a cohesive narrative structure.

02 Visual + Narrative

I used Midjourney to create all the image assets for the zine, while I focused on the layout, photo editing, and visual language. I want the publication to be carefully curated. Since the target audience is human, all information should be communicated in an effective, personal way.

EXHIBITION: SCREENPRINTING

The Art of Imperfection

When I think of generative AI, I see both sides of the story — black and white, strengths and flaws, process and results. I thought a screenprinted poster is the perfect medium to express this conflict. Just like how raw and imperfect screenprinting is, GenAI has its inherent variations and glitches, but it is filled with layers to uncover and endless possibilities to explore.

For the poster design, I applied a halftone treatment to an AI-generated image of a robot. The final poster is printed in silver and white ink on black paper.

EXHIBITION: THERMOCHROMIC INK

Seeing and Unseeing

To capture the ambiguity of generative AI, I incorporated an interactive layer into the exhibition. I extracted the most frequently mentioned keywords from the interviews and printed them on a sheet of poster paper. Above this, I placed a thin acrylic layer coated in thermochromic ink. When viewers place their hands on the surface, their heat will dissolve the black pigment, revealing the underlying keywords and issues related to GenAI that are often overlooked.

FINAL DESIGN

PIXELATED

The final zine was typeset in Redaction 35 and Suisse Int’l, saddle stitched with 9 signatures, 8 pages for each signature.

The exhibition took place on June 4th, 2024 at the University of Washington's Masters of Design Studio.

Why do you design?

To promote the exhibition, my classmate Gerri and I worked on the visual identity and marketing of the show. We designed an 11x17 poster and a series of Instagram stories that offered a sneak peek into each of the 5 projects. The poster was distributed across campus, and the stories were shared on @_uwdesign.

REFLECTION

On GenAI

My conversations with designers revealed a wide range of opinions on GenAI. Some are highly enthusiastic about its potential and possibilities, while others remain doubtful and wary of its use. The process of putting together this issue and exploring GenAI tools solidified my view that GenAI is a very powerful tool, but cannot yet replicate the unique style and taste of human designers. With this new technology comes a new set of responsibilities, but human control remains paramount.

On this quarter,
behind the scenes

This project challenged me in so many different ways. It was my first time systematically coding data from design research, my first print publication from start to finish, my first screenprinting, my first design exhibition, my first time pulling three all nighters in a row, and my first time being in such a small studio class with less than 10 people.

I’m very grateful for how much I learned in just 3 months. If I had more time, I would:
‍

Conduct primary research with more GenAI experts to answer the issues and concerns raised by fellow designers

Do more research when creating the project timeline to more accurately assess the time commitment for each section

Prioritize completion over perfection

This was a very rewarding journey in my college career, and I appreciate you reading to the very end :)
‍