AI, Mobility and Cities

About AI, Mobility and Cities

Welcome to the Mobility Forum, where the leading minds in transportation come together to shape how we move. About Me: zhaojinhua.com Mobility Forum Newsletter signup: https://zhaojinhua.com/mf/

Publisher

Jinhua Zhao

Category

education

Language

en

Explicit

No

First episode

31 May 2025

Latest episode

10 Oct 2026

Latest episodes

17 episodes in the feed.

1. 

   10 Oct 2026

   Grounded AI in State DOTs: What California, Texas, and Utah Are Learning

   What does it actually take to make AI work inside a transportation agency? In this episode, Jinhua brings together AI leaders from Caltrans, Texas DOT, and Utah DOT to explore how state transportation department are moving from AI experimentation to implementation. TxDOT is saving money through AI-powered invoice processing. Caltrans is exploring AI-enabled traffic management as an alternative to costly infrastructure investments. Utah DOT is building practical AI tools while investing in workforce training. Beyond individual applications, the discussion examines the harder questions: How should agencies choose AI use cases? When must humans retain control? How do organizations measure public value, manage risk, and scale AI responsibly? Making AI work in DOTs requires more than better technology. It requires better workflows, reliable data, human accountability, and organizational leadership.

2. 

   6 Oct 2026

   ​From Building Models to Judging Results: Fantastic AI Frontier in Transportation Research

   AI in transportation research has gone through four eras: pre-DNN models, deep neural networks, large language models, and now AI agents. Each one changed what AI could do. Each one also changed the role of our theory. In the pre-DNN era, the analyst wrote the model: the variables, the equations, the form of the utility function. Data estimated a relatively small number of parameters. In the DNN era, the model learned its own features and relationships from data; the analyst chose the data, the architecture and what to predict. In the LLM era, one model pretrained on a vast body of text and code could take on many tasks; the analyst described the task in plain language, often with a few examples, and supplied the data. In the agentic AI era, AI plans and carries out multi-step work: it writes the code, runs the models and reports the results. The analyst sets the goal, reviews the steps and judges the answer. Each era also changed the role of our theory. In the pre-DNN era, it was the center of our work. Deep learning and language models pushed it to the side. Now it is coming back. The AI methods that work best build established theory into the model, such as discrete choice, network equilibrium and conservation of traffic flow, and use data to learn what the theory misses. They tend to hold up better than pure AI on rare events, shocks and new policies. Theory is both the structure of these models and the check on their answers. Theory gives them structure. It also lets us check their answers. Seven frontiers are worth watching. Six are specific to transportation: Generative world models for mobility systems. AI for behavioral, social, and multi-agent simulation. Multimodal foundation models for transportation. AI for decision, control, and large-scale optimization. Trustworthy and verifiable AI. AI for transportation science. One is general. 7. Recursive self-improvement. AI that improves its own code and training could speed up every frontier above. Three of the most innovative young scholars in the field will take them on and join me for a conversation. Featuring... Shenhao Wang, University of Florida, on AI and behavioral models. a) AI does not replace classical behavioral models. It adds behavioral realism: satellite imagery captures the built environment that shapes travel decisions but that numeric variables miss. b) it generalizes them: models built on graphs relax restrictive assumptions while keeping classical models, such as the logit, as special cases. Lijun Sun, McGill University, on a) Seeing the invisible: AI as inference, not just prediction. Most quantities that transportation decisions actually need are never observed. b) From error bars on forecasts to error bars on decisions. We should report uncertainty on the promised improvement itself. c) Transportation uncertainty is endogenous: the decision changes the world it was optimized for. Sharon Di, Columbia University, on AI for transportation science and trustworthy AI. a) AI should move beyond prediction toward scientific discovery: which latent conflict patterns precede crashes, and under what conditions congestion pricing improves both efficiency and equity. AI can propose hypotheses, but transportation science must test them. b) in a safety-critical system, a model that works 99% of the time can still fail in exactly the situations that matter most: rare crashes, construction zones, extreme weather, unexpected human decisions. First, uncertainty-aware AI. Second, causally grounded AI. Third, verifiable AI.

3. 

   28 Sept 2026

   80% of Heavy Trucks Can't See Cyclists in Bike Lanes: Direct Vision, Blind Zones, and Truck Safety Laws

   The trucks on our streets are getting larger, their blind zones are increasing, and the risk of more — and more serious - crashes are on the rise. But what can a city or a state do? Massachusetts is in the midst of reckoning with this issue. Through the implementation of a first-in-nation type law on truck safety devices and the largest study of medium and heavy-duty truck direct vision in the country we are starting to find solutions to this challenge. 15 years of collaboration between Massachusetts municipalities, the state, and researchers at USDOT Volpe to eliminate fatal crashes between trucks and cyclists on our roadways. Featuring... The disproportionate negative safety impacts large trucks have in fatal crashes in MA The path of a city ordinance into a statewide law on side guards and cross-over mirrors The findings of MassDOT's direct vision study with USDOT Volpe Some potential pathways for what is next in the Commonwealth Guests: Kris Carter, the Chief Innovation Officer at the Massachusetts Department of Transportation and the co-founder of the Lab @ MassDOT, a small team focused on prototypes that scale to improve the ways people safely move about Massachusetts. In past lives he ran the Mayor's Office of New Urban Mechanics at the City of Boston, served as an advisor on transportation policy to a few different Mayors, taught physical education in a public elementary school, produced documentaries, and spent time as a wildland fire fighter. He’s also responsible for countless bad ideas, some failed experiments, and more than a few hopes that were crushed by bureaucracy. Two commentators bring the wider picture. Dr. Alex Epstein, Chief of the Vehicle Regulations at the United Nations Economic Commission for Europe and formerly at Volpe, was the researcher behind the Boston and Massachusetts studies; he will place the Massachusetts approach against the UN regulation on direct vision for heavy vehicles that Europe now applies to new trucks. Becky Mueller of the Insurance Institute for Highway Safety built much of the blind zone measurement method the study adapted; her work on passenger vehicles shows that as blind zones grow, pedestrian crashes grow with them, and IIHS is now extending the measurements to Class 3 pickups and cargo vans.

4. 

   19 Sept 2026

   What Does It Take to Operate a Driverless Fleet? Inside the Infrastructure of AV Fleet Operations with Ming Maa, CEO of Moove

   What does it take to keep a fleet of driverless vehicles running 24/7? Much of the autonomous vehicle conversation focuses on the technology inside the car. But once AV begin operating at scale, an entirely different set of challenges emerges: Where do they charge? Who cleans and maintains them? How do operators manage hundreds of vehicles across multiple facilities? Where does the electricity come from? And what must cities do to prepare? In this episode of Mobility Forum, Jinhua Zhao speaks with Ming Ma, co-founder and CEO of Moove AV, which operates Waymo fleet in Phoenix, Miami, and Las Vegas. Ming takes us behind the scenes of autonomous mobility—from charging, maintenance, sensor cleaning, parking, and fleet orchestration to power constraints, workforce needs, weather, deadheading, and the economics of operating robotaxis at scale. The autonomous vehicle itself may account for only 30–40% of its total five-year cost of operation. The remaining 60–70% comes from the operational ecosystem around it. The conversation also explores what this transition means for cities. Power availability, zoning, parking, pickup and drop-off design, and coordination with AV operators could increasingly determine how quickly and how effectively autonomous mobility scales. The autonomous vehicle may drive itself. But the system around it does not operate itself.

5. 

   19 Sept 2026

   The Next Generation of Activity-Based Travel Demand Modeling with Michel Bierlaire

   EPFL's Michel Bierlaire — creator of Biogeme and one of the field's foundational voices in discrete choice modeling — joins Jinhua Zhao to lay out the next generation of travel demand models. The shift: from individual trips to full-day activity schedules from an assumed decision hierarchy to every decision on the same level, with the data deciding the order from isolated travelers to households, colleagues, and social networks from a static population snapshot to synthetic populations that live whole lives Michel explains: how to manage the choice set explosion in four moves: Formulate → Optimize → Choose → Estimate: billions of possible daily schedules written as constraints, solved without enumeration, simulated for random utility, and estimated from a sampled choice set. why the payoff is the broader set of questions a model can pose, not forecast precision: telework as a change in the primary activity, who wins and who loses below the segment average, trip chaining, school hours, off-peak rail demand how a synthetic population acquires a life course: demographic priors updated against each census what happened when the Swiss Federal Railways tested the method inside its own activity-based model The question for transportation agencies is not whether the latest activity based model is more precise. It is whether it allows us to ask—and answer—questions we could not before.

6. 

   2 Aug 2026

   Canada Carries Twice the Transit Ridership Per Capita: What Vancouver and Toronto Do That US Cities Can Learn

   Metro Vancouver: 153 transit trips per resident. Metro Boston: 54. Calgary: ~77. Denver: ~21. The cities look remarkably similar. Their transit outcomes don't. In this episode of the Mobility Forum, I spoke with Jeff Busby (COO, TransLink), Josh Colle (Chief Strategy & Customer Experience Officer, Toronto Transit Commission), and David Cooper (Leading Mobility Consulting) about Canada's experience: Better transit doesn't always require bigger infrastructure. More often, it comes from better operational decisions. Four takeaways: Better service, not cheaper service, has been Canada's biggest ridership advantage. A painted bus lane can demonstrate demand long before a multi-billion-dollar rail project opens. Protect bus speed. Congestion is an operating cost, not just a customer inconvenience. Signal priority and dedicated lanes often deliver some of the highest returns on investment. Fund transit as regional infrastructure. Transit creates value beyond its riders—from reduced congestion to stronger local economies—so its funding should reflect those broader benefits. The lessons are Canadian in origin, but highly relevant to agencies across the United States.

7. 

   26 Jun 2026

   Gas Tax Is Going to Zero, Oregon DOT Spent 25 Years Building the Replacement

   In November 2025, Oregon made a per-mile road usage charge mandatory for EVs and hybrids — 2.3 cents a mile, starting July 2027. This episode of Mobility Forum, Jim Whitty, who began the work in 2001, and Frederic Charlier, CEO of ClearRoad told us how it finally happened. If we want to joint live session of Mobility Forum, subscribe here: https://zhaojinhua.com/mobility-forum/

8. 

   22 Jun 2026

   The Fiscal Cliff Playbook: Inside Chicago and San Francisco's Transit Reset with Tom McKone (CTA) & Julie Kirschbaum (SFMTA)

   Chicago and San Francisco turned a funding emergency into their fastest service reforms in decades. The frequent-bus, transit-lane, and headway-management moves you can borrow — plus the revenue strategy underneath them. Tom McKone, CFO of the Chicago Transit Authority, and Julie Kirschbaum, who runs the San Francisco Municipal Transportation Agency, run the most ambitious service reforms their agencies have attempted in decades. My conversation with them covers: The cliff is structural and not a passing dip, you can't cut your way off it, the service-quality playbook working right now in Chicago and San Francisco, the revenue fight that decides whether any of it lasts. If we want to joint live session of Mobility Forum, subscribe here: https://zhaojinhua.com/mobility-forum/

9. 

   14 Jun 2026

   Microtransit at Scale: Via CEO Daniel Ramot on Citywide Transit and the AI in the Control Room

   How a DOT with no transit history, no appetite for risk, and a skeptical public built a microtransit system that works ? Arlington, Texas with 400,000 people, 99 square miles, 3 failed transit votes over 25 years— built citywide public transit for a fraction of the cost of conventional fixed-route transit. A ride is now available within a two-block walk of every resident. I sits down with Daniel Ramot, CEO of Via, and Alicia Winkelblech, Director of Transportation for the City of Arlington, Texas, who lived the transition firsthand. Via now serves 800+ cities across 30+ countries, powering microtransit, paratransit, school bus, transit planning, and autonomous-vehicle networks, and went public on the NYSE in 2025. This episode is for the practitioner staring at a bus network bleeding both ridership and budget. It walks Arlington's route from the inside: how to procure for outcomes instead of vehicles; how to de-risk a launch so a skeptical public will tolerate it; and whether on-demand transit can scale to meet the needs of growing cities, and what role it plays alongside fixed-route systems. We also covers the role of AI in the next generation of public transit. How Via is using it to reshape planning, scheduling, and real-time decision-making for DOT and cities.

10. 

    11 Jun 2026

    Stop Benchmarking Copenhagen: The Cities You Should Be Learning From Instead

    Adam Millard-Ball mapped walking and cycling across 11,587 cities in 121 countries. The results may surprise you: density and bike networks matter far more than climate, walking delivers most of the emissions gains, and your best peer city is probably one you've never studied. If you want to join us live at Mobility Forum or receive updates, subscribe at ⁠⁠zhaojinhua.com/newsletter⁠⁠ (http://zhaojinhua.com/newsletter)

11. 

    11 Jun 2026

    The Robotaxis Arrived Before the Rules: AV Lessons From San Francisco, Austin, and Washington DC

    Today’s episode is my conversations with three city leaders sitting at three different phases of AV deployment: Rachel Castignoli in Austin, Tilly Chang in San Francisco, and Stephanie Dock in Washington, D.C. We’ll walk each city in turn — the Wild West, the rulebook, and the capital. Three cities, three phases of deployment — and the same three gaps: no data, no first-responder protocol, and no authority to fix either. If you want to join us live at Mobility Forum or receive updates, subscribe at ⁠zhaojinhua.com/newsletter⁠ (http://zhaojinhua.com/newsletter)

12. 

    11 Jun 2026

    Don't Govern AVs the Way We Governed Uber: Heaven, Hell and Who Decides

    Today's episode is my conversation with Kara Kockelman and Robin Chase. With AVs, the technology won't decide whether a city gets Heaven or Hell. The governance will. Here's the lever cities finally have, the empty miles towatch, and the pricing fight worth having. If you want to join us live at Mobility Forum or receive updates, subscribe at zhaojinhua.com/newsletter (http://zhaojinhua.com/newsletter)

13. 

    11 Jun 2026

    Research Stack Powering the World's Most Experienced Autonomous Driver

    As of Nov 2025, Waymo is delivering over 250,000 rides a week. In today's episode, I sit down with Drago Anguelov Head of Research at Waymo to unpack the AI behind autonomous driving: from 3D Sensing & Spatial Reasoning to AI foundation models and Social Intelligence. If you want to join us live at Mobility Forum or receive updates, subscribe at ⁠zhaojinhua.com/newsletter⁠ (http://zhaojinhua.com/newsletter)

14. 

    10 Jun 2026

    Inside Transport for London's Secret Sauce: How TfL Got London’s Businesses to Fund a £19B Railway, and the 4 Moves US Cities Can Steal

    What does it take to build a world-class transportation system - and keep it world-class for 25 years? I am Jinhua Zhao, Welcome to the Mobility Forum - where the leading minds in transportation come together to shape how cities move. Today, I sit down with Shashi Verma, Chief Technology Officer of Transport for London - the man who has personally touched every major innovation TfL has delivered over the past two decades. from Oyster Card, contactless payment to Crossrail and congestion charging. If you want to join us live at Mobility Forum or receive updates, subscribe at ⁠zhaojinhua.com/newsletter⁠ (http://zhaojinhua.com/newsletter)

15. 

    17 Jun 2025

    Ecosystem Alignments: What China’s EV Surge Teaches the U.S.

    Industry experts analyze China's EV surge, revealing that ecosystem alignment, not just government control or subsidies, drives their success, offering key lessons for the U.S.

16. 

    2 Jun 2025

    AV for Cities #02: If cities are not yet prepared for AVs, how should they?

    Industry experts explore how cities can strategically leverage their power and policy instruments to actively shape the future of autonomous vehicle integration for enhanced mobility.

17. 

    31 May 2025

    AV for Cities #01 Are Cities Prepared for Autonomous Vehicles?

    Industry experts discuss how cities are largely unprepared for autonomous vehicles despite advancements in the technology, emphasizing the need for proactive planning and engagement.

Reach and audience

Public platform figures. Ratings count people who left a rating, not total listeners.

Apple Podcasts (US)

5.0 / 5

2 ratings

Spotify

5.0 / 5

3 ratings

Podcast Authority Score: 15 / 100

A composite of feed quality, social presence, YouTube performance and engagement. Read the methodology.

Quality

31

Social presence

0

YouTube

0

Engagement

0