By Dr. Lance B. Eliot, the AI Insider for AI Trends and a regular contributor
Senator, you’re no Jack Kennedy. That’s a famous line that was uttered by Senator Lloyd Bentsen during the Vice-Presidential candidate debate against Senator Dan Quayle in 1988. During the debate, Dan Quayle had somewhat equated himself to the heroics and stature of former President Jack Kennedy. Lloyd Bentsen struck back with the biting words that as someone that knew Jack Kennedy, and given the achievements of Jack Kennedy, in no manner whatsoever could Dan Quayle be considered the equivalent. It was a remark that served to cut down Dan Quayle and vaulted Lloyd Bentsen into the famous quotes stratosphere. The same kind of line is used today in both everyday conversation and in political stump speeches, and used to cut down an opponent during a discussion or outright debate. Having incredible stickiness and memorable as a slur, anyone today needs to be cautious in trying to compare themselves to something or someone else that is great, because your opponent can trot out a variation of the “you’re no” and do the contemporary drop-the-microphone to win the debate.
Why do I dredge up this notable line? Because there are some that want to equate the recent emergence of self-driving shuttle vans to the emergence of AI self-driving cars. By-and-large, this is a misleading comparison. These self-driving shuttle vans are usually a shallow version of what a true self-driving car is supposed to be and do. I’ll explain why the comparison is so misleading. Worse too, it has the danger of confusing the public about the nature of self-driving cars. In the same breath, it will actually have some positive impacts on self-driving cars and so I’ll cover that too. Overall, though, I’d like to clearly state this: I know all about AI self-driving cars, I have been writing all about AI self-driving cars, and I can say that the emerging self-driving shuttles are “not Jack Kennedy” as in they are not AI self-driving cars.
A recent rumor is that Walt Disney World is going to deploy self-driving shuttle vans at their resort properties in Florida. It is said that either or both of Local Motors (Phoenix) and Navya (Paris) might supply the shuttles. Disney will presumably do an initial series of tests, likely using “cast members” (Disney employees) as the initial passengers in the vehicles. If you’ve ever been to the massive properties that Disney has in Florida, you well know the need to have shuttles that take you throughout their resort locations. Distances between their theme parks, restaurants, hotels, and other venues can be in the dozens of miles. Already there is a small army of shuttles, buses, boats, and all manner of transport provided. There several monorails too, along with bike rentals and rickshaw modes of transportation. Thousands upon thousands of guests and employees are in constant motion inside and throughout the properties. It is like a massive ant farm.
Currently, the manpower needed to undertake all this transportation is enormous. Zillions of human drivers are needed. They need to be trained in how to drive the vehicles. They need to be available and be at their assigned vehicles when needed. They need to be in their right-mind and not distracted or somehow say drunk or otherwise unable to properly drive. Some will be sick on given day or shift, or play hooky and not show-up for work. Substitutes need to be ready to step-in. All those drivers need to be supervised and monitored. They need to be responsive to changes in schedules and changes in routes. At times, they interact with guests and so need to showcase themselves in a Disneyesque way of being pleasant and friendly.
In short, from a management perspective, it’s a nightmare. The logistics is sizable and it is a daily chore of ensuring that things go right. If transportation goes sour, the guests go sour. If the guests go sour, the resort loses the guests. If the resort loses the guests, the revenue drops. The transportation is not something that you might at first think is crucial to the success of the Disney parks there, but it is. Transporting people is one of those seemingly hidden aspects that until you have a transportation glitch or stoppage you just take it for granted. The Disney locations in Orlando are especially vulnerable since they are so spread out. In contrast, the Disneyland resort in Anaheim, California is much more compact and the transportation aspects are not nearly as vital to the ongoing operations there.
Without a doubt, it makes a lot of sense for Disney to want to use self-driving shuttles at their Florida parks. No human drivers means that you can wipe out all the troubles and costs of having those human drivers. No need to worry whether the driver will show-up for their shift. No need to worry whether the driver is alert and aware while driving the guests. Consider all the labor costs of regular time and overtime that is needed to keep the transportation going. Knock down that cost and the savings are impressive. It is pretty much a no-brainer that if you could get rid of the drivers you would be better off in nearly all respects.
I know you might say that the drivers are human ambassadors and that guests are enchanted by their smiles and pleasantries. Yes and no. Getting a human driver to be smiling and pleasant is a hard thing to do. They might hate their job and take it out on the guests. They might have had a bad day and are taking it out on the guests. They might get verbally abused by a guest and then take it out on other guests. There is no guarantee that the drivers are going to be good human ambassadors. Meanwhile, if the robotic shuttle has some kind of pre-recorded super-friendly voices that emit messages and funny anecdotes, the odds are that guests will as enchanted as they would be a human driver, and also that Disney will avoid the dangers of those human drivers that treat the guests poorly.
Think about this aspect of using self-driving shuttles and you realize that Disney can also capitalize on their Tomorrowland type atmosphere. People love to ride on the monorails because it gives them a sense of what the future is supposed to be. Remember when you were a child and the entire world was supposed to eventually have those slick, safe, and clean monorails? Sorry, we are still not there. But, anyway, guests going to the parks will likely love to ride on the self-driving shuttles. It is another added “magical ride” at the resorts and will portray Disney as once again an innovator that is trying to paint a bright and better future. I am betting that acceptance by guests will be high, the costs for providing the transportation will drop, and even likely the transportation will run more smoothly overall. The loss of jobs for the human drivers will hurt, and there will likely be attempts by labor to curtail or stop the onslaught of the self-driving shuttles, yet it is unlikely that the affected labor can muster enough of a protest to prevent the conversion.
Let’s get back to my original premise about the nature of these self-driving shuttles. I wanted to emphasize that they are not the same as AI self-driving cars. Someday in the future they might be, but the ones that are rolling out in the foreseeable future are pretty much brainless. First, we’ll consider the simple physical aspects of the shuttles. A typical self-driving shuttle holds about a dozen or so passengers. It is usually equipped with the same kinds of sensory devices that you see on a self-driving car, including LIDAR, cameras, conventional radar, etc. Seems on the surface like it is just like a self-driving car, other than maybe being a bit larger since it has to accommodate a dozen or so passengers. They are often shaped rather tall to allow passengers to get in and out without having to stoop.
What is the difference then between the self-driving shuttles and the AI self-driving cars? It’s all about where they drive, how they drive, and the “thinking” or AI involved in the driving. These self-driving shuttles are usually confined to a designated geographical area. They are zoned into a geo-fence, which limits where they can go. This is important because it keeps them in driving areas that are well-known and constrained. A true AI self-driving car can go anywhere, and does not need to be limited to a particular geographical area. The shuttles are kept in a cage.
Furthermore, the cage is mapped and mapped again. The shuttle “knows” entirely the routes and what the roadways consist of. In the case of AI self-driving cars, a Level 5 (topmost capability) self-driving car must be able to handle new roads and new routes, and be able to figure out what to do on-the-fly. For the shuttles, it’s almost entirely pre-canned. There isn’t much computational effort involved in terms of gauging what the road ahead might be and how to navigate it. Even though you can’t see any apparent rails, think of the shuttles like trains on rails, and a train just moves ahead on a pre-determinate path. That’s what the shuttles do.
Speed is another important difference. The shuttles usually go about 5 to 10 mph, moving along at a slow pace. This is helpful to the shuttles since they have time to use their sensors to figure out whether a guest has stepped into the roadway and thus come to a halt. Halting is pretty easy at that speed. Though the shuttles can usually go faster such as 20 to 30 mph, doing so is risky. The stopping time is longer and so less safe. Plus, stopping at that speed will likely toss around the guests that are riding in the shuttle. And the sensory data needs to be processed much more quickly.
A true AI self-driving car is expected to zoom along at freeway speeds of 70 to 80 mph. The sensory data needs to be rapidly obtained and processed. Decisions about driving by the AI needs to occur extremely rapidly. Actions taken will need to take into account the long stopping times and the dangers of swerving or turning upside down the car. I think we all instinctively know that going 5 mph versus 80 mph is a big, big, big difference. For those of you that have teenagers that drive, you probably started them in a parking lot where they drove at 5 mph, and you dreaded the fact that soon they would be on regular roads at 20 mph and then eventually on freeways at 80 mph. The difference of driving skills and the time that you have to make decisions and react is like the difference between night and day between those slow speeds and the faster speeds.
I think you can see why I say that the self-driving shuttles aren’t Jack Kennedy. They are kept in a geographical cage, the cage is well mapped beforehand, they tend to take the same routes over and over, they move at slow speeds, they can stop readily easily, and they don’t need to do much in terms of cognitive types of skills. Self-driving cars need to be able to navigate all kinds of roads and routes, they are not constrained within a geo-fence, and they need to be able to have cognitive skills equivalent to humans in order to properly drive the vehicle.
This is not to say that Disney is somehow mistaken by going ahead with the self-driving shuttles. I think it makes a lot of sense. I am just saying that the reporters that then declare that we have arrived at the era of self-driving vehicles will be mistaken and will be providing “fake news” (see my recent column about AI fake news). Over and again, we’ll likely see gushing stories by reporters that are wide-eyed and excited to see the future has arrived. Look, they’ll say, proof positive that self-driving vehicles are ready for the roads. I am sure that Disney will allow reporters to go on the shuttles when first trials have occurred, and those reporters will dutifully file stories about the wonderment of the shuttles.
In that sense, it will be “good” because it will add more fuel to the energy and drive toward true self-driving cars. The stocks of firms that make self-driving cars will get a boost, since the feel-good stories about the self-driving shuttles will spread over to the self-driving cars arena. Very few will realize that these are night and day differences. All they will know is that a vehicle with four wheels and transporting passengers seemed to be able to work and do so without incident. Voila, self-driving cars must be good too.
There is a dual-edged sword to this. Suppose that a Disney self-driving shuttle gets into an accident. Yikes! This could drive down the stocks of self-driving car makers. If Disney cannot make it work, who can? Some pundits will step forward and point out that Disney is not a Google, it is not an Uber, it is not a Ford. How could one expect a company that makes themes parks and fun family-friendly movies to also be at the forefront of self-driving vehicles?
Anyway, as you can see, the advent of self-driving shuttles such as those that are possibly coming to Disney will be a means for the public and regulators to get a taste of self-driving vehicles. It won’t be the same as self-driving cars and it won’t be using as sophisticated AI, but at least it will be using the rudiments and allow for further progress in the direction of true self-driving cars. The self-driving shuttle companies realize they need to continually up their game, and that they cannot allow their shuttles to remain in the brainless category for very long. Once self-driving cars get good, the expectations for the self-driving shuttles will rise too.
You might be aware of the children’s story about the little train that could, which involves a train that keeps saying “I think I can” as it tries to proceed ahead. The self-driving shuttles, even as brainless and robotic as they might be at day one, are like those trains that are striving to someday be thinkers. Making and selling them today for highly constrained circumstances is handy and a money maker as society warms up to the self-driving future. People might not know that they are not yet in the future when they ride on one of those shuttles, but at the same time they are supporting ultimately a future involving self-driving cars. Those brainless shuttles will someday get their brains. Don’t be surprised if you get onto one of those shuttles and hear the song from the Wizard of Oz about if it only had a brain. It will soon enough.
This content is original to AI Trends.