Uncovering the hidden labor powering artificial intelligence (AI) through research, design, and embodied experiences.
WHAT
Design Master's Capstone
WHEN
2024
Opportunity
AI is often presented to the public as a fully autonomous and almost magical tool. Symbols like the sparkle emoji (✨) have become synonymous with its presence in everyday technology, making it seem more approachable. Yet, this framing can downplay the complex realities behind its development, particularly the essential work of data workers who meticulously label and annotate the data that fuels AI systems. By illuminating these unseen human contributions and offering an immersive, hands-on experience, we can challenge conventional views of AI and spark meaningful discussions about the ethical implications and hidden costs.
OUTCOME
Artisanal AI, an interactive art exhibit showcased at the 2024 ACM Conference on Fairness, Accountability, and Transparency (ACM FAccT), immersed visitors in a world of data work. The exhibit featured an office-like space with task manuals and laptops featuring real-world data labeling tasks, such as gender classification, satellite image annotation, and sentiment analysis of e-commerce data. As visitors worked through these tasks, the manual prompted them to reflect on what it means to give data meaning. After completing the tasks, they shared their thoughts on a collective reflection wall. In collaboration with the Data Workers Inquiry, the exhibit also showcased curated stories from data workers, offering personal accounts of the individuals behind AI systems.
ROLE + PROCESS
I conducted qualitative design research to examine the hidden labor and power dynamics within the AI ecosystem. This began with secondary research on data worker exploitation, followed by firsthand experience working on platforms like Amazon Mechanical Turk. I also interviewed a range of stakeholders—including data workers, AI researchers, product managers at data vending companies, venture capitalists, data labor organizers, and policymakers. These interviews revealed a shared cognitive dissonance about the opaque nature of the system, making it difficult to see the parts and the whole. Across all actors, there was a collective pressure to build AI quickly. These insights led to the design of an intervention that would be visceral and thought-provoking. I facilitated co-design sessions with data workers, labor organizers, and AI researchers from DAIR and Turkpoticon to create Artisanal AI. This exhibit aimed to bring visibility to the often-overlooked human labor behind AI systems and provoke critical conversations about the ethical implications of AI development.
Curious about the full process? Feel free to get in touch!
IMPACT
