Tashkent, Uzbekistan (UzDaily.uz) — The most valuable transport asset of the coming decade may not be another road, railway or port. It may be the intelligence systems already flowing through the existing infrastructure. Across Asia and the Pacific, artificial intelligence (AI) is already making transport safer, faster and more reliable without a metre of new construction.
The shift is underway across every mode, starting with urban roads. In Singapore, intelligent transport systems sit at the heart of the Smart Mobility 2030 strategy: cameras detect accidents and breakdowns so response teams can be sent quickly, while roadside signs redirect drivers before congestion spreads. Rather than expanding capacity, Singapore is making better use of the roads it has. In the Republic of Korea (ROK), computer vision detects road damage that once required manual inspection. Autonomous micro-pods carry elderly residents on first- and last-mile trips, and autonomous buses run on fixed overnight services, extending inclusive mobility to low-density routes that are otherwise uneconomic.
Railways are moving in the same direction. Delhi Metro is expanding computer vision to improve passenger safety in stations, while Australia’s trains fitted with sensors turn routine journeys into continuous track inspections, flagging maintenance needs without dedicated inspection vehicles.
Ports and maritime shipping are also changing. Autonomous vessels in China and the Russian Federation are improving navigation and reducing waiting times, while generative AI at terminals in ROK handles documentation for dangerous goods and assesses container damage. None of this required a new port; it simply makes existing infrastructure perform better.
In each case, the technology was only part of the story. Introducing AI into operations involves more than acquiring new software. Four conditions decide whether it works. The first is infrastructure. AI in transport runs on reliable electricity, digital connectivity, computing capacity, trusted local data and effective transport systems. Second, adaptation is more important than importation. A model trained on traffic conditions in one city may not perform well in another city where road conditions, vehicle types and commuting patterns differ.
The third is people and institutions. AI literacy, reskilling and institutional capacity decide whether a pilot project becomes a public service. The fourth is cooperation across borders. Shared data standards, joint research and technology transfer let smaller economies reach capabilities no country can build alone.
Those conditions matter now because the old formula is running out of road. For decades, when demand grew, Governments across Asia and the Pacific built more roads, railways, ports and airports. Those investments connected communities and supported economic growth, but cities are now more congested, infrastructure costs more to build and maintain, and climate pressures are intensifying. By 2050, two-thirds of the world's population is expected to live in cities. With transport emissions rising and the region facing a transport investment gap in the trillions, new infrastructure will still be needed, but building alone can no longer keep pace with these pressures.
AI helps because infrastructure—built to move people and goods—generates data on how networks are used and assets perform. Rather than a single technology, AI is a family of them: computer vision reads camera images to detect road damage or safety risks; machine learning forecasts demand and predicts which assets need maintenance; language models turn regulations, maintenance records and incident reports into guidance.
None of these works alone. Sensors collect the data. Networks carry them to models for analysis. Traffic centres, logistics platforms and operators turn the findings into action. The value lies in the link between data, digital infrastructure and human judgement.
The risk is that this capability spreads unevenly. According to AI Preparedness Index of IMF, Asia and the Pacific includes both world's most AI-ready and least prepared economies. In 2023, 65 per cent of developed economies in the region already adopted national AI strategies, but the rest did not have national strategies in place. In 2023, 65 per cent of developed economies in the region already adopted national AI strategies, but the rest did not have national strategies in place yet. Many are entering the transition with no plan for skills, infrastructure or risk governance.
AI in transport is still at an early stage of deployment. According to an OECD study, transport is among the top quarter of sectors for new AI inventions, measured by patents, but among the bottom quarter for the share of companies actually using AI. In other words, the technology is being developed much faster than it is being put to work and there is still time to shape how AI is used in transport, and who benefits from it.
The United Nations Decade of Sustainable Transport (2026–2035) recognises AI as an enabler of safer, more efficient and sustainable transport. Technology alone will not deliver that. The real work is ensuring every country can adopt, adapt and benefit from AI, so the transformation narrows development gaps rather than widens them.
The next transport revolution may not be measured by how much more we build, but by how much more value we unlock from what already exists, and by whether every country in Asia and the Pacific can shape and benefit from it.
---
Authors:
Katrin Luger, Chief of Transport Research and Policy Section, ESCAP
Francesco Spandonaro, Intern, Transport Division, ESCAP
