The people and data behind AI
A research group at the intersection of human-computer interaction and machine learning. We study the people and data that fuel powerful AI systems, and work to ensure that AI is used responsibly.
Motivation
We conduct research at the intersection of human-computer interaction and machine learning. Our research projects investigate questions related to the people and data that fuel powerful “artificial intelligence” technologies.
Each data record upstream of an AI system was created with the involvement of a person (and more often than not, a group of people), so many “data-centred” projects involve people and many “human-centred” projects involve data. Our work seeks to understand these relationships and empower people to control how their data is used, and ultimately, ensure that AI is used responsibly.
Work
Data Levers
You are a crucial source of data needed to train and test AI. You can use your data flow as a bargaining chip; send data to AI systems you support, and don't send it to those you don't.
Data Napkin
This interactive tool allows you to adjust assumptions and play with the numbers. The underlying math is simple arithmetic, yet it provides a framework to reason about order-of-magnitude estimates.
AI Text Similarity Playground
An interactive playground that explores similarity-based measures of AI influence on writing. It accompanies the research paper "Overreliance in Writing Tasks: Exploring Similarity-Based Measures of AI Influence on Writing and Proposing a Reflective Writing Interface Intervention".
PeasantSoon
Coding-agent transcripts capture your entire engineering process, yet today the AI companies — not you — control that trove. A data contributor-oriented platform that keeps agent traces alongside the code they produced, with fine-grained access controls and redaction so you decide who sees them and can claim your share.
Latest publications
View all publications- Mechanism Plausibility in Generative Agent-Based Modeling
Patrick Zhao, David Huu Pham, Nicholas Vincent
FAccT
- An Audit and Analysis of LLM-Assisted Health Misinformation Jailbreaks Against LLMs
Ayana Hussain, Patrick Zhao, Nicholas Vincent
AIES
