Full time
Harvard University
Remote (Cambridge, MA, USA)
The Institutional Data Initiative (IDI) is a new research center working to advance society’s relationship with knowledge by expanding access to, and deepening our understanding of, the data that underpins AI. By collaborating with library, government, and academic institutions to publish their knowledge collections as AI training sets, IDI seeks to 1) empower those institutions and the cultures they represent, 2) build a foundational pipeline for academic inquiry of AI, and 3) advance the state of the art for all builders of AI systems. IDI’s work spans the AI data ecosystem—from digitization, data structuring, and metadata synthesis, to safety and security analysis, all the way through to benchmarking and the development of ethical and governance frameworks. Institutional collaboration forms the gateway to this work and IDI places a particular emphasis on opportunities with institutions that expand the cultural breadth of knowledge represented in the building blocks of AI. At...