Project lead: Dan Sinykin (Emory, Department of English)

Project team: Nia Judelson (Emory, Ph.D. ’26, English), Em Nordling (Emory, Ph.D. ’26, English)

Artificial Intelligence is powered by data. When that data is private and opaque, as it is in the AI initiatives at major tech companies, we are prevented from knowing the full ethical and political implications of the resulting AI. The AIAI Data Collective is a new project dedicated to modeling transparent and ethical AI data collection, documentation, storage, and sharing. We will complete an environmental scan and literature review to produce a white paper on the state of AI data. From that white paper, we will produce a set of best practices for AI data. We will build a website to house and share potential AI training data according to these best practices, modeling what ethical AI data looks like.

The AIAI Data Collective applies Critical AI frameworks to latent questions of AI data storage in order to best store the datasets produced by other AIAI projects. To do this, we are currently developing a digital tool that will guide users through the decision-making process for ethical AI data management. Methodologically interdisciplinary, this project pulls from technical fields and library sciences, with a particular emphasis on humanities orientations. Specifically, we apply frameworks such as feminist theory, critical race theory, and other social, cultural, and economic frames to the issue of AI data storage. This tool will ultimately encourage users to consider the labor practices, sustainability, privacy, and representational issues that underlie AI data storage.