RethinkAI

Generative AI has fueled growing interest in identifying use cases that can improve business and government efficiency. But in the civic sector, efforts to put AI to work have lagged. RethinkAI is a collaboration between New America, the Urban Tech Hub, the Burnes Center for Social Change at Northeastern University, and the Center for Media Innovation & Social Impact at Boston University. This consortium aims to fill this gap.

RethinkAI works along three paths that form a continuous loop of doing and learning:

• First, build out ambitious local pilots. Rooted in local experience and expertise, the team brings together public sector officials, community partners, and AI experts to co-design pilots that explore the full potential of AI.
• On a parallel track, construct a transparent learning and research agenda. A robust research program will guide local experiments; openly report what worked, what didn’t, and why; and translate that knowledge into guidance for practitioners everywhere.
• In turn, translate local pilots and research into new practice and national-level policy. Through regular knowledge dissemination, empower critical audiences—including national policy makers, technologists, and philanthropists—to implement a new practice for public problem solving.

In 2025, the Urban Tech Hub partnered with one of the nation’s first microgranting organizations, Citizens Committee for New York City (CitizensNYC), to explore the opportunity for AI in the civic sector. Urban Tech Hub research staff and Cornell Tech urban tech students supported CitizensNYC leadership and staff through several activities:

• Spring 2025 Public Interest Technology (PiTech) studio course. Students interviewed CitizensNYC staff and constructed user journeys for a proposal analysis support tool and mocked up a dashboard demonstrating how open city data could be leveraged to provide additional context during annual microgrant proposal review.

• Summer of AI. Urban Tech Hub staff and student researchers conducted three workshops with CitizensNYC staff, developing working prototypes for three use cases identified in collaboration with the team—the automated generation of grant snapshots for board review packages, automated web search for due diligence research prior to award disbursement, and automated lookup of electoral districts for awarded grants to support in fundraising outreach and engagement.

• Proposal Analysis Dashboard Pilot. In Fall 2025, Cornell Tech and CitizensNYC are pursuing the creation of a proposal analysis dashboard, building off earlier collaborations. This tool will leverage multi-modal large language models as well as geospatial foundation models to combine applicant-provided materials with other contextual data to help identify trends in need over time and space and draw out semantically-related themes not easily identified at scale through manual review.