MIT Transit Lab to develop open-source AI platform for transit agencies

The MIT Transit Lab will develop the Public Transit Intelligence Hub, an open-source platform designed to bring public transit monitoring, operations control and passenger communication into one system. Google.org is providing $2.1 million, and the project was among 15 selected for the organization’s Impact Challenge: AI for Government Innovation. The funding covers a three-year project; Google.org will also make engineers and AI product experts available on a pro bono basis.

Transit control centers track information from radio feeds and computer screens, including camera views, vehicle locations, riders, traffic and road conditions. The researchers say those inputs are fragmented rather than combined in a system that presents an overall view of network conditions. That environment leaves staff making operational and communication decisions that can affect passengers who rely on transit.

PTIQ is intended to provide a centralized, AI-orchestrated system that helps control-center staff make better-informed decisions and gives riders more immediate, accurate information. Its decision-support interface is expected to bring together predictive models, optimization engines and contextual reasoning based on large language models. The project is not designed to make operational decisions on staff’s behalf; transit employees will remain responsible for weighing the options.

The project’s co-principal investigators are Awad Abdelhalim, associate director of the Transit Lab, and Jinhua Zhao, head of MIT’s Department of Urban Studies and Planning. MIT Lecturer Jim Aloisi will manage the program, which will involve the Transit Research Consortium and researchers from the Transit Lab, the MIT Mobility Initiative and Northeastern University. Google.org’s support therefore includes both funding and technical assistance from its own staff.

For AI builders, the proposal puts information integration and human decision-making at the center of an operational AI system. The researchers describe transit operations as dynamic and involving multiple stakeholders, with many tasks lacking one objectively correct answer. They also emphasize that AI needs to work within an agency’s organization and earn staff trust, rather than being judged only by model benchmarks. PTIQ’s stated approach is to put machine-generated analysis alongside human judgment while agency staff retain authority over decisions.