Week 6: Mapping Our Values in the Age of AI
Reading Discussions
-
How does an image become data that a machine can learn from?
-
What role do categories and labels play in training an AI system?
-
What kinds of decisions have to be made when building a dataset?
-
How might different ways of organizing the same images lead to different outcomes?
-
What does these two videos help you notice about the relationship between people, data, and machine learning?
Guest Lecture and Workshop
Tuan Huang is a PhD candidate at Eindhoven University of Technology. Her research lies at the intersection of design, education, and AI, exploring the values that emerge between these fields. Using Research through Design (RtD) methods, she examines how design practice can shape AI and how AI is reshaping design practice.
Her recent work, “Workmanship of Learning: Embedding Craftsmanship Values in AI-Integrated Educational Tools,” presented at ACM CHI, explores how AI can support agency, creative struggle, and productive friction in design education.
The session has two parts. First, we will explore the alignments, misalignments, and tensions between our values as designers and those embedded in existing AI tools. Through collage and annotation, we will map our identities and practices onto the current technological landscape, using aesthetic expression as a form of critical inquiry. In the second half, Huang will share her work and discuss her broader research practice around design and AI values.
Readings
Assignment 6
This assignment is part 1 of the project Uncertain Archive.
Submit Assignment
Step 1: Collect
Begin with something you can recognize before you can fully explain it: a feeling, gesture, visual tendency, social behavior, kind of object, recurring situation, way of speaking, family habit, neighborhood detail, design language, or other phenomenon that keeps catching your attention.
Build an archive of at least 100 items around it. The archive can contain photographs, screenshots, text, drawings, sounds, objects, recordings, found material, fragments of conversations, your own work, or anything else that belongs to the world you are trying to understand. The archive should be a reflection of your own observations and experiences, not a replication of existing datasets.
Step 2: Map
Using this week’s reading on Morphology, Taxonomy, and Typology, organize the archive into a map. You do not need to follow only one method. Try different ways of making relationships visible and notice what each structure allows you to see.
You might organize through:
- morphology: qualities, features, parts, variations, or transformations
- taxonomy: categories, hierarchies, families, and distinctions
- typology: recurring types, forms, behaviors, or recognizable patterns
Your map does not need to resolve the archive into a perfect system. Look for places where the system becomes unstable: an item that belongs to several groups, a category that becomes too broad, an unexpected repetition, a missing type, or a relationship that you can sense but cannot yet name.
Bring Next Week
- 100+ items
- one or more maps of the archive
- the organizing logic(s) you used
- a few patterns, exceptions, contradictions, or uncertainties that emerged