If you’re looking forward to the day that R2-D2 is your car’s pilot, sorry, it is not going to happen. Yes, you can read, watch a movie, or work on some contracts before you get to the office, all while you whiz along. We will undoubtedly have cars that handle the navigational details, where you are merely “the cargo” as a passenger. It is inevitable, and not even that far in the future.
However, we are probably not going to see a wasted passenger space for a robot body when it will all fit on a single circuit board tucked away somewhere in the vehicle. Doing away with the pedals, steering wheel, and other controls will provide a lot more room and comfort.
The greatest likelihood is that many of us won’t even own cars. What is the point? Artificial Intelligence (AI) managed self-driving car services can have an appropriately sized vehicle in your driveway in just minutes. Why would you want to own one? Pay-per-use means everything is built into the cost of hiring. No more insurance; no more fuel (they’ll probably be electric anyway); no more maintenance; no burdensome expenses.
We’ll continue to sell vehicles to individuals for several years. This would be particularly true of specialty vehicles (green cars, pickup trucks for small businesses, off road, adventure types, or camping vehicles), but an AI acquainted with current trends could show that movement away from personal vehicles has already begun. Manufacturers should be looking toward fleet sales in the coming years as personal sales decrease.
In the future, the cost of insurance, if you have a self-drive option, would probably be outrageous anyway, even if the AI did 99% of the driving—that 1% would represent a substantial risk. Driving is a skill that needs to be exercised regularly to maintain competence—infrequent driving would erode that ability.
It will get to the point that ordinary members of the public won’t trust human drivers. People make too many mistakes. No, for practicality’s sake, let the Googles and the Ubers fight it out for your business; let them buy the huge fleets that are necessary to keep people moving in our society, and which, under fleet-administration, run very economically.
Insurance risk can currently be optionally assessed by data modules supplied by insurers. They use those to determine if people commit unsafe acts, or if they are excellent drivers that deserve a significant rate cut. AI would provide a much richer report (40 mph in a school zone between 8-9 am or 3-4:30 pm; heavy acceleration away from every traffic light; constant lane changing and failure to signal, etc.) recording excellent skills, or chronic errors in judgment and arrive at a customized rate for each driver.
There’s more to it
Naturally, Tesla™ is all gung-ho about self-driving cars (to the extent that all of their vehicles manufactured since 2016 have the installed hardware, ready-to-go when it is legally approved). Google™ is equally enthusiastic. Other forward-looking companies (e.g., Uber™, that fully expect to have a driverless fleet by 2021), who are currently throwing millions at the idea regarding both research and development, are still relatively rare.
Most companies are adding AI features a little bit at a time. They select items that won’t intimidate a dedicated driver; they choose things that are more like enhanced safety features, which are easier to accept. These “features” will handle inadvertent lane departures, emergency braking (if something or someone enters the path of the vehicle), or Collision Avoidance Systems for when another vehicle stops abruptly.
Some cars now come equipped with radar in the rear bumpers to tell you when other vehicles are in “blind spots,” with audible and visual warning systems when you operate lane-change indicators or turn signals. They also keep you from reversing over children and bicycles in your driveway by automatically applying the brakes. More sophisticated systems even manage to parallel park your car, if you somehow managed to get a driver’s license without learning that skill (Video: How is that possible?).
AIs can Build Cars, too
When we think of car-building robots, we equate that with AI, but it is an entirely different thing. Automated manufacturing is programmed via computer coding. It has no interpretive skill; it merely repeats an action over and over again.
Preparing an AI requires little programming skill. Such a procedure is littered with terms like Neural Nets, Deep Learning, Machine Learning, and so on. Ultimately it boils down to a robot/program being slaved to a human performing the desired task. By following, and recording actions and techniques, it learns how to deal with a missing part, an out-of-place component, a bad weld, or a stripped bolt.
It observes the human in action, learning the subtleties, recognizing deviations, and seeing how such differences are resolved. It can be a direct-slave, the same as the method used to create CGI effects in movies, or it can review months or years of video and data in a very short time.
It doesn’t even need to interact with people. If it is a process-slave, it can watch how a balance-sheet is reconciled—how problems are solved—and gain its expertise through observation. That means that AI isn’t stuck on the factory floor; an AI System can function anywhere where there is repetitive work that requires an interpretive ability but is too menial for a human. That displaced person can then handle more diverse or creative tasks that the AI would encounter difficulties solving.
Until we get fully automated, AI can make self-driving safer. It can monitor the driver’s reaction times, eye-movements, and response to traffic situations; it can gauge the operator’s response to speech, report medical emergencies, and even bring the vehicle to a safe stop if the situation warrants.
However, far from being alone and solely responsible for interpreting the entire world around it, AI-powered vehicles would share data continually. A vehicle, even if not part of the network, would be observed, and its actions would be integrated into the behavior of the surrounding vehicles.
In practical terms, if a self-driven vehicle was approaching a red light without slowing down, cross traffic would be informed and could take actions to avoid a collision. Such vehicles, recklessly endangering the public, would be automatically reported to the authorities, and perhaps even remotely disabled.
In the world of cars without drivers, there would also be cars without passengers moving to places where they were needed. These “empties” could be used to isolate and stop a misbehaving self-driver without endangering law officers.
Building a three-dimensional map of your surroundings is certainly not the easiest task in the world; it baffles a lot of human drivers (see the video above). AIs, on the other hand, can interpret data at thousands of times per second, much faster than a human being can do so.
Some truly amazing transport truck drivers can swing at a 44-foot long trailer into a dock, dead-straight, in one attempt. Conversely, there are those that could never accomplish the task, no matter how many attempts they were permitted.
With an AI running that show, and with cameras on every corner of the vehicle, every parking job would be perfect the first time. In traffic, such vehicles would be flawless, too. Nobody would fall asleep; trailers could drive continuously from source to destination, meaning products would arrive faster, produce would arrive fresher, and costs would be lower.
You would never read about dump trucks tearing down electrical transmission wires all along a street, or hitting bridges because their dump boxes were still raised. No more jack-knifed trailers, or tipped trailers that were taking corners too fast. Driving would become utterly dull, and ideally suited to letting AIs handle the job.
Clients hate being put “on hold.” Numerous studies have shown that a caller made to wait more than 15 to 20 seconds has increased cortisol levels and stress. It’s a very bad way to start an interaction with a customer.
As AIs get more sophisticated, they will be able to communicate directly with customers, even over the phone, and troubleshoot complex problems. The speech emulation can be so successful as to be undetectable to the caller. AIs are perfect active-listeners, remembering everything the customer has said before and relating it to new information.
Until that is commonplace, we have ChatBots that are completely capable of dealing with basic customer inquiries. Customers are very willing to go online because they like the instant response of a ChatBot, and the fact that they don’t have to wait for a human to become available—people nowadays are very sensitive about how their precious time is consumed.
ChatBots can handle the booking of appointments for car-servicing at dealerships, explain the various types of service, provide cost and time estimates, and if the inquiries exceed their capabilities, connect clients with a human.
An onboard AI can monitor the performance of the car, optimizing how it functions. By reporting its status, the manufacturer can roll out software updates that improve the function of the vehicle, even performing remote maintenance, and reminding drivers when it’s time to get some scheduled service that requires a physical visit. Manufacturers will also get lots of feedback about real-world use, and learn how to make improvements.
And of course, as we have seen with the OnStar™ system, vehicles can self-report emergency situations, if the driver is unresponsive after, say, an airbag deployment. An AI could identify someone not permitted to operate the car and disable it, or report vandalism, or attempted theft, even if unattended.
In an area with poor communication coverage, the AI network, where vehicles communicate directly with each other, could pass along its emergency information to any passing car. When that vehicle reaches an area with a proper connection, the message would be passed along the network.
Data Mining is an essential process to all modern business. AIs use Neural Networks, which are specifically designed to mimic the ways human brains store information and access it. Whether in real life or a computer simulation (the AI wouldn’t know the difference), it can perform a task millions of times, trying many variations, to find preferred solutions and to eliminate dead-ends.
When integrated with the knowledge of hundreds or thousands of human experts providing solutions to problems, the AI becomes an Expert System, capable of combining the knowledge to create a unique answer to a problem. This lends itself to being able to model trends, to identify increasing customer desires before the clients themselves notice. It means improved logistics, sourcing materials, scheduling production, optimizing design, and maximizing consumer desire through improved marketing schemes.
This predictive/interpretive ability also means more after-market sales & service, and loyalty building. Fleets will need telematics to monitor driver performance, efficiencies, or errors. Even once they are fully automated, it will still monitor maintenance, current vehicle status, as well as location, traffic conditions, and timetables. Establishing good relations now you’re your clients means that they’ll think of you first when such options become de rigueur.
It has been said that that Artificial Intelligence is the last great invention that humanity will need to create. There will always be people that want to perform practical research and expand human knowledge. The totality of humanity’s learning, from prehistory up until now, is half of what we will know in 18 months. That’s right: we double our knowledge every 1½ years.
We will keep assembling information because we’re curious species. The challenge is that we cannot analyze it as fast as we collect it. We may already possess the knowledge to do things that are currently unimaginable, but it is too far-flung, in too many different disciplines, and we might never pull all the pieces together on our own.
If, however, we turned an AI loose on our data, it is the ideal candidate to find all these obscure associations. Anti-gravity, Warp Drive, Teleportation… what do you think is “impossible”? The smart bet is that AI will back prove you wrong…
If you want to learn more, we provide webinars to show you where the future is going and how you can get there. Don’t shortchange yourself. Join us today!