MIT Transit Lab receives $2.1M to build an AI platform for public transport

MIT Transit Lab has received $2.1 million from Google.org to develop the Public Transit Intelligence Hub, or PTIQ. Selected as one of 15 projects in the global AI for Government Innovation challenge, the three-year effort aims to create an open-source platform for public transport control centres.

PTIQ is intended to connect systems that often operate separately today: real-time network monitoring, operations control, and passenger communications. Control centres receive radio feeds, camera data, vehicle locations, station activity, passenger information, traffic conditions, and road reports, but fragmented systems can make it difficult to form a unified view of the network.

Decision support rather than staff replacement

The planned decision-support interface will combine predictive models, optimisation engines, and contextual reasoning based on large language models. The project team stresses that the system is not intended to automate operational decisions. Awad Abdelhalim, the project's technical lead and a co-principal investigator, says the goal is to ensure that human decision-makers have the best information available.

That distinction matters because public transport operations are dynamic, involve multiple stakeholders, and rarely offer a single objectively correct response. The researchers intend to leave the assessment of competing options to transit employees while using AI to organise incoming information and suggest possible responses.

Expected benefits and current limits

The team expects PTIQ to improve response times, reduce crowding at platforms and bus stops, and provide passengers with more timely information. These are proposed benefits of a system still under development, however, and its real-world impact will need to be established through deployment and evaluation.

Staff trust and organisational fit are also central to the project. Jinhua Zhao, the other co-principal investigator, argues that the question is not only whether AI can perform a task, but whether it can function inside the organisation and earn employees' trust. Alongside funding, Google.org will provide pro bono assistance from engineers and AI product specialists during the three-year project.