Honda R&D President Yoshiyuki Matsumoto talks about cars and AI with SoftBank technology. (Image courtesy of NIKKEI Asian Review.)
Now, thanks to the Internet of Things (IoT) as well as inventiveness from SoftBank and Honda, this dream might come true.
The two organizations will be working together to merge SoftBank’s human--like robot Pepper with the entertainment system in Honda cars. In the style of Pepper’s current functionality, the car will be able to address the driver’s talking while assessing data from sensors and cameras to gauge the driver’s emotions.
The announcement comes after SoftBank has agreed to acquire IoT chip designer ARM Holdings for $31 billion. No doubt these chips will play a vital role in the new IoT AI car.
Pepper is a bit of an IoT technical marvel and has become an instant hit in Japan. It currently is helping in stores and in the health-care industry.
But does Pepper have a place as your backseat driver?
Why Do Cars Need Artificial Intelligence?
Replica of K.I.T.T. from Knight Rider at Universal Studios. (Image courtesy of Wikipedia user Morn the Gorn.)
He has a point. One of the key questions any engineering design team has to ask itself is, ”How will adding IoT functionality improve the user experience for a product?“
In the case of a car, much of this is already known: GPS, satellite radio and hands-free calling.
So the question is, will adding AI to this mix really improve the user experience?
Yes, it would. Sometimes.
The short answer is that having an artificial driving companion can be useful for those long drives. It can help keep you awake. Additionally, as the car can assess your mood, it can also help keep you happy. It can also offer information and assistance when needed. All of these benefits will help keep the roads safer. And the technology is reminiscent of our old friend K.I.T.T. from Knight Rider:
K.I.T.T.: Michael, I detect a certain tone in your voice.
Michael Knight: Oh, what tone is that?
K.I.T.T.: The one where you're convinced you've come up with a brilliant idea. The next thing know, you're doing something foolhardy and in my surveillance mode.
Michael Knight: K.I.T.T.?
K.I.T.T.: Yes, Michael?
Michael Knight: Go into your surveillance mode.
One of the biggest design flaws with many of the current IoT functionalities incorporated into cars is that they can be very distracting and infuriating. This doesn’t make for a good user experience. In theory, adding an AI can help improve these issues.
For instance, you can call your family hands-free as your car connects to your phone. However, it will take five attempts before the computer will recognize if you want to call your wife Tina or sister-in-law Nina to see who can pick up the kids.
However, the AI might be able to figure this all out for you via IoT connections and big data analytics. By checking both Tina and Nina’s schedules and cross-referencing with the timing of the school bell, the car can call the right person for you before you even remember to ask.
The same can be said for that pesky GPS. The majority of GPS systems are still not hands-free. Having an AI to change the destination on your navigator based on appointments, phone calls and big data analytics will be a lot safer than having to fumble with the device using your hands.
The Downside of an AI Car on the IoT?
“Open the car garage door, HAL!” “I'm sorry, Dave. I'm afraid I can't do that.” (Image of HAL 9000 from 2001: A Space Odyssey courtesy of Wikipedia user Cryteria.)
Additionally, human-AI interaction is very tricky. If the personality of the car isn’t quite right, you could end up cruising the uncanny valley.
Combining an intelligent computer with all of your life data can then lead to a frustrating argument with the vehicle. What happens when the big data analytics tells the car one thing and you tell it another? Say, for example, a garage door is locked and you can’t remember the password.
“Open the car garage door, HAL!”
“I'm sorry, Dave. I'm afraid I can't do that.”
Will Driverless Cars Have AI?
“This is a recording. I'm afraid Holly is busy at the moment, but if you would like to leave a message after the bleep, he will get back to you. Bleep.” (Image of Holly from Red Dwarf courtesy of Wikipedia.)
In a sense, this would bring the human race full circle with the loving steeds of old.
However, relying on the IoT for driverless cars can present a potential issues. What if the connection goes down? Or what if there is a denial-of-service (DoS) error? An automated message from a self-driving car’s AI will provide little comfort when you are traveling at highway speeds.
Such was the case when the computer in the sidesplitting scene from the Red Dwarf was moving at light speed and his AI was too taxed to answer any of the crew’s questions:
“This is a recording. I'm afraid Holly is busy at the moment, but if you would like to leave a message after the bleep, he will get back to you. Bleep.”
Another question is, what happens if the sensors are fooled and the driver is too distracted by the AI to notice? A similar scenario appears to have occurred in the tragic Tesla Autopilot crash, where later reports hinted that a running video of Harry Potter might have been involved.
These are the types of scenarios that all engineering design teams must ask themselves about when adding IoT functionality to their products. Otherwise, you might end up with some IoT designs that have laughable or dangerous concepts like the ones in this previous article.
What are your thoughts on AI cars? Would you trust your life to the computer of the Enterprise when traversing an asteroid field that is sucking energy reserves dry, making all the systems unreliable? Comment below.