In the times we are in, it’s interesting to chart the course of AI-inspired transportation technologies, and try to predict when fully autonomous vehicles will be here.
You can read about existing programs at flagship companies like GM and Penske, or turn to academia. For example, there’s this report from MIT Sloan, which indicates that AI will soon automate $65 billion in transportation work, making previously labor-intensive processes a lot easier, and unlocking efficiencies in getting people and cargo where they need to be.
“Today, the transportation industry has reached a remarkable stage where vehicles can operate without any human intervention, thanks to cutting-edge technology,” writes Hardik Shah at Prismetric, outlining the forecasted growth of the sector in the years to come. “These improvements have been quite helpful in bringing about new ideas and changes in the field. We are currently at a time when AI is changing transportation in big ways that are getting the attention of industry executives all around the world.”
But when AI comes to transportation, as it currently is, it will come to land, air and sea, too.
Pondering AI-Age Transport
At a recent event at Stanford, a panel of experts talked about what they are most excited about in terms of AI for transportation. You can watch the full video of the conversation here.
Panelist Ernestine Fu Mak, co-director of Stanford FTL, cited advances in mobility and aviation, noting her involvement in a company called Shield AI, that is working on intelligent vehicle design for defense and other industries.
“(They’re) looking at autonomy for aviation, from drones all the way up to fighter jets. They have some really great technology around dogfighting, swarming, … and on the ground transportation side, for instance, which is where I initially made investments, in autonomous vehicles.”
Panelist Marco Pavone, an associate professor at Stanford, noted his enthusiasm for pushing the boundaries of autonomy across various projects, and of the acceleration of development processes.
He also mentioned foundation models.
“First of all, the foundation model is a very hyped model these days,” he said, attempting to describe this technology in detail. “For me, a foundation model is a model that has been trained on internet-scale knowledge … typically in a self-supervised fashion. It can then be adapted to many different types. Such a definition is very broad. So there are many different types of foundation models, like video generation models, vision foundation models … those models that are capable of producing human-like visual interfaces. So in other words, they’re capable of taking a problem step by step, and explaining .”
Evgeni Gousev of Qualcomm said he is excited about advances in silicon, citing examples like a Snapdragon ride pilot program, and the development of hybrid AI systems.
Was America Falling Behind?
Panelist Sampriti Bhattacharyya, who has worked on projects like flying boats, painted a rather dark picture of transportation competitiveness in America, saying the U.S. is “number 11” in maritime, and noting the absence of what she called “marching fleets,” and a fall-off in ship building over recent decades.
AI, she noted, takes labor-intensive practices like ship-building, and makes them orders of magnitude easier.
“We will have a lot more manufacturing in the U.S.,” she said, of AI’s impact more broadly.
Big Changes
Later, the group talked about innovations like the use of synthetic data, and simulations like digital twinning, and efficiencies.
“If you think about AI and AI applications through ground transportation, if you think beyond just the impact it has on the individual vehicle, and making that autonomous, but (think more broadly) from a network perspective, and how AI can be used to automate traffic routes, you know, the broader, broader network effects of that, I think that has huge implications on sustainability,” Fu Mak said, suggesting that right now, our human-driven cars sit idle around 95% of the time. “Of course, I think there’s a little bit of efficiency paradox … I do think we’re in a really interesting time when it relates to energy and sustainability. In some ways, I think AI is indirectly solving our energy crisis, just by the amount of capital that all of these big tech companies are now putting into nuclear reactors.”
Pavone noted the work that is going on at Stanford’s Sustainable Mobility Center, and went over various facets of how scientists approach AI for transportation.
“The four main facets that we are focusing on are environmental sustainability,” he said. “In terms of electrification, the power train, alternative fuels, for example, aviation impact on the energy grid… then urban infrastructure, so how do we think (about) urban design … safety … and then equitability and financial sustainability.”
Highlighting main objectives in this vertical, Bhattacharyya suggested that future vehicles should be all-electric, autonomous, and composed to scale.
“Sustainability needs to be addressed in a holistic way,” Gousev added. “Look at the system holistically. One part of the system is the energy efficiency of the hardware. That’s basically your horsepower, right? And this is happening. The other way is, how do you design a system that is efficient.”
Does all of this ring true as we get more self-driving cars on the American road? I think so. If you’re reading this, there’s a good chance that you have ridden in one of these next-gen vehicles yourself. What did you think? And how long before this kind of thing becomes the norm in your neighborhood? At the same time, we’re likely to see AI transform defense and aviation too, pretty soon. And don’t forget its capabilities at sea.
