Graduate Studio:
Technology B

Week 1–4: Crack Inside the Black Box

This module examines how intelligence has been modeled, simulated, and represented across different historical and cultural contexts. Through examples such as the Buddhist concept of the five skandhas, the I Ching's rule-based systems of interpretation, Ramon Llull's combinatorial thinking machine, Alan Turing's Imitation Game, John Searle's Chinese Room argument, and Ned Block's Blockhead thought experiment, students will examine how intelligence has been defined, modeled, and contested across cultural, philosophical, and computational contexts.

Rather than treating artificial intelligence as a recent technological breakthrough, the module situates contemporary AI within a longer history of attempts to formalize cognition, prediction, and decision-making. With critical readings, students will be challenged to compare how different systems define intelligence, what forms of knowledge they privilege, and how their underlying assumptions shape human perception and behavior.

By unpacking these historical precedents, students develop a critical foundation for understanding contemporary AI systems not as magical autonomous intelligences, but as designed structures that emerge from datasets, rules, interfaces, and human decisions.

Project 1: Epistemological Research of Technology

Students will select a historical or computational model of intelligence and reconstruct its underlying logic through research and prototyping. Working with simple rules, decision trees, conversational scripts, classification systems, cellular automata, or other procedural frameworks, students will create a functioning artifact that demonstrates how the model produces the appearance of intelligence.

Possible precedents include Joseph Weizenbaum’s ELIZA, Claude Shannon’s mechanical mouse Theseus, Marvin Minsky’s SNARC, Markov chain, John Conway’s cellular automata, or rule-based expert systems such as MYCIN. Students may also draw from artistic precedents such as Jean Tinguely’s Homage to New York, the Useless Machine, or Sun Yuan and Peng Yu’s Can’t Help Myself, considering how artists have used automation, feedback, and procedural behavior to question what counts as intelligence.

Alongside reconstructing these systems, students are encouraged to consider how their prototype might simulate not only reasoning, but also affect. They may explore whether a minimal system can produce a convincing sense of emotional response or relational presence. Rather than passing a traditional Turing Test based on correctness, the project opens the question of whether a system can pass a more subjective or emotional threshold, where users perceive intention, personality, or care.

Week 5–9: The Coded Gaze

This module introduces machine learning as a shift from general artificial intelligence and rule-based programming to systems specifically trained on classification, prediction, and decision-making. The algorithms powering today's software and generative tools actively scrape, label, and flatten everyday visual culture. This coded gaze inherits the legacy of Enlightenment-era taxonomy and colonial record-keeping, and now relies on the invisible human labor of data labeling to enforce existing social hierarchies and oversimplify nuanced complexity.

Drawing on Ruha Benjamin’s New Jim Code and Cathy O’Neil’s Weapons of Math Destruction, we will examine how algorithmic bias manifests in real-world systems. Case studies such as predictive policing and automated hiring reveal how these technologies often reinforce structural inequalities and produce homogenized social outcomes. The same underlying logics are embedded in the generative tools designers increasingly rely on, where they risk flattening aesthetic diversity and cultural specificity into standardized, repetitive outputs.

In response, contemporary practices in interaction design and creative coding actively interrogate and expose these embedded biases. We will study projects such as Mimi Onuoha’s The Library of Missing Datasets, Trevor Paglen and Kate Crawford’s ImageNet Roulette, Joy Buolamwini’s Gender Shades, Zach Blas’s Facial Weaponization Suite, and Caroline Sinders’ Feminist Data Set. Together, these works demonstrate how art and design can function as critical methodologies, offering speculative approaches, auditing techniques, and alternative frameworks that challenge the systemic flattening of AI slop.

Project 2: Uncertain Archive

In Project 2, students will act as both archivist and algorithm to examine how classification systems generate content, shape knowledge, and reproduce power. Students will manually construct a corpus of 100 to 500 distinct items and organize it using Morphology, Taxonomy, and Typology (MTT), approaching these systems as subjective and interpretive frameworks rather than neutral structures.

Using accessible tools such as Teachable Machine, Runway, ml5.js, or image recognition platforms, students may ask a machine to label, sort, transform, or extend their dataset. This process can include classification, clustering, tagging, or generative techniques such as style transfer or synthetic image production. Rather than aiming for technical precision, the goal is to observe how the dataset shifts when interpreted or re-produced through an automated system.

Through this comparison, students will begin to identify what each system makes visible or invisible. Manual classification may foreground nuance, intention, and situated knowledge, while AI systems often prioritize pattern recognition, scale, and speed. At the same time, both approaches introduce forms of bias, omission, and distortion. Students are encouraged to treat these differences not as problems to resolve, but as material to work with.

The final outcome will translate these observations into a design form. Using patterns, sequences, and a chosen medium, students will construct a work that reflects the relationship between human judgment and machine processing. A short written reflection will accompany the project, considering how both systems produce meaning, where they break down, and what kinds of cultural or aesthetic assumptions they reinforce.

Week 9–15: Disrupting the Chain-of-Thought

This module narrows our focus to Large Language Models (LLMs) and the contemporary obsession with natural language processing. Operating entirely on statistical probability, commercial systems like ChatGPT and Google NotebookLM simulate genuine comprehension through massive textual extraction. We will analyze how various proprietary and open-source models differ in their architecture, yet similarly inherit histories of industrial efficiency, cybernetics, and techno-utopian thought. By reducing collective cultural memory to computable weights, these foundation models frame intelligence as an extractable resource. Consequently, the prompt interface and techniques like chain-of-thought reasoning act as control mechanisms that enforce a rigid, standardized logic on how knowledge is produced.

Grounding our critique in contemporary theory, we will challenge the narratives of seamless automation pushed by the technology sector. Foundational texts from Emily M. Bender, Timnit Gebru, and Kate Crawford expose the ecological devastation and invisible ghost labor required to sustain these massive models. To move theory into practice, students will learn how to run and design with local, private LLMs on their own machines. Taking ownership of the model to bypass corporate infrastructure, prioritizing privacy, environmental care, and localized knowledge over computational scale.

Finally, the module explores how art and design can prototype speculative systems to subvert these dominant language paradigms. Students will examine work from scholars such as Allison Parrish, Stephanie Dinkins, and Suzanne Kite, alongside artists like Rashaad Newsome, X.A. Li, and Lauren Lee McCarthy. By analyzing these interventions, students will ultimately reclaim these language technologies as sites for critical world-building and design alternative interfaces that actively embrace uncertainty, slowness, and interdependence.

Project 3: Alternative Now(s)

For the final project, students will create a speculative design prototype that reimagines artificial intelligence beyond dominant narratives of productivity, optimization, and accelerationism. Using tools explored throughout the semester, including local language models, custom datasets, archives, physical computing, publishing, or other experimental media, you will imagine an alternate reality grounded in different values and ways of knowing.

This project asks you to move beyond pure critique alone. What happens after the critique is written? How do we avoid both technological utopianism and dystopianism, leaving space for desire, ambiguity, care, and collective imagination?

Rather than asking how AI can become more efficient or intelligent, consider how computational systems might support memory, ritual, ecological relationships, uncertainty, or other forms of knowledge generation. Projects such as Stephanie Dinkins's Not The Only One, Allison Parrish's computational poetry; Bina48, a robot built to simulate the memories and personality of an African American woman; Suzanne Kite's integration of Lakota epistemologies into digital systems all offer examples of how computation can become a site of cultural, political, and speculative inquiry rather than technical optimization.

Your final outcome should be exploratory, open-ended, and less didactic than the current existing scenarios. Whether it takes the form of an interface, publication, installation, artifact, performance, or something else entirely, the medium is up to you. We are not looking for solutions or predictions of the future, but for thoughtful propositions that expand what technology could be. By sitting with the tensions and possibilities of these tools, you will explore how art and design can offer a glimpse into reclaiming computation as a site of imagination, cultural meaning, and alternative world-building.

Your performance will be evaluated on its own merits, not based on comparing your work with other students. We focus on your understanding of concepts and your ability to apply them in a meaningful way. We don’t prioritize technology efficiency or professionalism as the primary criteria. We value the learning process and recognize that everyone progresses at their own pace.

Should you have any inquiries regarding grading, please don't hesitate to reach out to the faculty directly.

Active participation is essential and comprises 25% of the final grade. This includes, but is not limited to: keeping up with readings, assignments, and projects, contributing meaningfully to class discussions, active participation in group work, and coming to class regularly on time.

While attendance is one aspect of active participation, absence from a significant portion of class can compromise successful attainment of the course objectives. We consider a significant portion to be three weeks or 20% of class time. Lateness or early departure from class may be recorded as one full absence. We encourage you to let us know as much in advance as possible if you need to miss a class, are running late, or need to leave class early. More than three uncommunicated absences, late arrivals, or early departures will result in a deduction from your final grade.