Preparing for an agentic era of human-machine transportation systems: Opportunities, challenges, and policy recommendations

Abstract

Human-Machine Transportation Systems (HMTS) refer to transportation systems where humans and machines interact to enable mobility. The history of humans creating and utilizing machines for transportation purposes dates back from the invention of the wheel to more recent innovations such as bicycles, automobiles, traffic signals, handheld navigation devices, asphalt pavers, and computer-aided design and management tools. In recent years, technological advancements have transformed machines from passive tools into more active participants, with humans increasingly delegating complex tasks and responsibilities to them. While these advancements have revolutionized mobility, their siloed, uncoordinated implementation has also introduced critical challenges, including urban sprawl, high fatalities, environmental degradation, and worsened societal disparity.

The recent advancements in machine learning, robotics, communication, and computing technologies prompt the emergence of agentic transportation systems (ATS) to potentially address these chronic issues and transform how people access resources and opportunities. In ATS, intelligent machines serve as autonomous intermediaries, facilitating the interactions among humans and between humans and infrastructure. From this new standpoint, early-stage ATS—such as autonomous vehicles, on-demand ridesharing platforms, generative design tools, construction robots, and anomaly detection equipment—have already begun to enter society, calling for an understanding about whether the current research and practice in transportation planning and engineering are ready for ATS. A review of recent literature reveals four main categories of research: (1) co-visioning, co-planning, and co-design; (2) co-construction and co-maintenance; (3) co-control, co-operation, and co-management; and (4) co-usage and co-consumption. The review suggests a significant lack of studies on the proactive integration of agentic machines within and across individual lifecycle phases, risking severe and irreversible consequences. Accordingly, the paper proposes a framework to guide the development of ATS to be justifiable, inclusive, and adaptable (JIA) and ensure the intelligence in and of the next-generation HMTS to be genuinely human-centered and societally beneficial.