UTTRI associated faculty Professor Baher Abdulhai took part in WRLDCTY 2021, a global forum for urban innovation, on October 27, 2021.
Professor Abdulhai was a panelist, along with Professor Igor Gilitschenski (Department of Computer Science at University of Toronto and Research Scientist at the Toyota Research Institute), in the session “Steering toward an AI movement future” moderated by Jan De Silva, President and CEO of the Toronto Region Board of Trade.
Professor Abulhai’s comments at “Steering toward an AI movement future”
Q: What does it mean for computers to “learn” and what applications does that have for your research?
Babies and infants learn to walk without a teacher, they get up and fall a few times. After a few bumps, they are under the bed and above the fridge, if you can only catch them. This is learning by direct interaction with the environment. Learning to reinforce what worked and avoid what did not. Can we make a traffic light learn that way? We let traffic lights “see” upcoming traffic using cameras and sensors and learn by trials how to time themselves to reduce waiting time. Have you ever seen a traffic light waste green time serving no one? We train them to learn not to do that. They become agile, allotting calculated times to each traffic movement without wasting any. They become smart. – Professor Baher Abdulhai
Q: Self-driving and autonomous vehicles could offer one of the most momentous transformations in urban life. What will these transformations look like and what roadblocks are we hitting to get there?
AVs can induce momentous transformation indeed, but be careful; transformation is a double edged sword. It can be very positive or very negative. AVs can drive safely in tight formations and hence increase road capacity, but if they are overly cautiously leaving space, they can reduce capacity which we do not want. They can be superhuman or subhuman, and of course we need to design them to be superhuman. Further, AVs are attractive. We all dream of taking a nap in the car or watch a movie without stressing about traffic jams. This can exacerbate dependence on the car as a mode of transportation. They can also encourage longer commutes and urban sprawl, in which cases cities will be in trouble. People embraced the car more than 100 years ago because those machines promised to take us farther, faster, in comfort and style. Then cities mushroomed. AVs repeat the same promise but on steroids, if we let them. But if they are used to complement public transit, first mile last mile, and if they are shared, then the urban impact can be positive. – Professor Baher Abdulhai
Q: Self-driving cars are one potential solution to our movement woes, but smarter traffic management is another. How does the implementation horizon of A.I. traffic management compare to autonomous vehicles? Are we facing challenges that differ from broader A.I. obstacles?
Making the car smart as in AV is necessary but not sufficient! There is the infrastructure dimension as well, and the space of public policy. Hopefully AI will be harnessed in three good ways to manage traffic: (1) complementing transit (policy), (2) make AVs achieve superhuman driving and hence expand the capacity of existing infrastructure, and (3) create smart infrastructure and smart vehicle-infrastructure interaction. On the last point, an AV cannot anticipate congestion ahead. A smart infrastructure can detect that congestion and not only inform upcoming AVs but can also use those upcoming AVs to pace traffic and avoid exacerbating congestion, or route them elsewhere. – Professor Baher Abdulhai
Q: There’s been a boom in e-commerce during the pandemic. E-commerce distributors around the region’s distribution districts saw demand peak – at one point 50% above 2019 levels. How can A.I. help to better move goods as well as people? Not just to the in final delivery but across the supply chain?
With such increase in last mile deliveries, AI can help in many ways. Smarter routing is one. AV sharing for delivery is another. Your car is typically parked at work all day. If it can go and make a chain of deliveries, this can save many trips by others, and save you parking as well. Robotic deliveries is also a rising domain, including on the ground robots and possibly flying robots. – Professor Baher Abdulhai
Q. Incongruous regulations are one of the greatest enemies of innovation and tech-adoption. How can governments better enable the scaling up of A.I. – especially as it pertains to transportation?
Innovation faces an eternal challenge: no one, especially governments, want to procure it first because it has never been done before!! Between innovation and implementation is a death-valley for which we need a bridge. Governments can build this bridge by investing in and accelerating field-testing and deployment of innovation, to close this gap. We must learn by doing, safe doing of course. – Professor Baher Abdulhai
Q: Building on this conversation about regulatory frameworks is data management and privacy. In Toronto and around the world, communities become distrustful of A.I. projects if their data is not being protected. As champions of smart cities, what can we do to better protect data and instill public confidence?
Data sharing is a double-edged sword. It is an enabler of innovation (look what you can do with simple Google maps). On the other hand, it can be privacy invasive. Look how much Google knows about you. More dangerously, data in the hand of hackers is a recipe for disaster: abuse of information, hacking your accounts, and maybe terrorism. The challenge is to find ways to encourage data sharing, while protecting people and assets. Heavy penalties and laws against data abuse are needed for protection. – Professor Baher Abdulhai
Q: What can our North American cities learn from Asia-Pacific economies about A.I. adoption? Are there any projects or regions you’ve had your eye on as particularly successful in this?
I would say the development, uptake and consumption of technology is much faster and more daring in many Asian economies. Not just Japan and Korea for example, but also China is of course a rising star as well. Look at where they are with 5G relative to the rest of world. They innovate in large numbers, and just by virtue of large numbers they will succeed. We need to pick up our pace to compete. – Professor Baher Abdulhai
Watch the session video recording
The session video recording is posted on UTTRI’s YouTube channel with permission of WRLDCTY 2021. Watch it here.
Access to all session recordings on WRLDCTY 2021’s streaming platform is free for registrants. Register using code TORONTOGUEST.
About WRLDCTY
WRLDCTY is an international forum (hosted in-person in New York, virtually everywhere else) taking place for the first time October 26 to 28, 2021. The goal is to bring together urban innovators and thought leaders to shape better cities for tomorrow. Themes range across urban planning, real estate and technology. Toronto, Vancouver, New York, Singapore, London, Houston and Hong Kong are the participating cities. These cities will each broadcast their own WRLDCTY sessions. Toronto’s four WRLDCTY sessions were hosted by Destination Toronto, Toronto Global, the City of Toronto and the Toronto Region Board of Trade.