Follow Datanami:
March 9, 2017

AI Controls Are Likely In Your Next Car

Kayla Matthews


Anyone who’s ever landed in the passenger seat with someone learning to drive will admit to being at least slightly nervous. Thanks to advances in artificial intelligence (AI) controls, you may never again have to worry if Jr. makes a mistake behind the wheel. Just make sure your car has enough miles under its belt to know how to avoid a wreck.

That’s right, artificial intelligence is coming to consumer vehicles, and it could happen as soon as five years from now. While the media has focused primarily on the way this new technology allows for potentially self-driving cars, there are a number of other features and perks that come along with it. Here are a few of them.

Toyota Aims to Reduce Accidents

You’ll be hard-pressed to find an automaker out there that doesn’t take an interest in building safe cars. But the advent of automated intelligence allows brands to take the concept a step further, by building cars that not only make it through a crash fairly well but also can actively avoid one.

In 2016, Toyota promised to deliver an autonomous car within the next five years, and it says it will be able to make adjustments in real-time to avoid a wreck.

The claim becomes more believable when you consider the number of automakers already producing technologies close to what the Japanese brand is suggesting. Many luxury marques now offer laser-guided cruise control that allows a car to accelerate or decelerate to match the speed of the driver in front of it. Tesla’s autopilot system even includes accident-avoidance features, though it recommends drivers never rely on the system by itself.

Toyota’s AI technology takes the concept a step further by allowing the car to be evasive beyond one lane. While it’s unquestionably a step forward, moral dilemmas remain the limiting factor in how autonomous vehicles make decisions about where to go in situations where a crash is unavoidable.

Nuance Controls Entertainment Through Voice Recognition

Automakers like Toyota are building intelligent cruise control into their vehicles (posteriori/Shutterstock)

Technologies like Apple CarPlay and Android Auto have given us a glimpse of the convenience we can look forward to from a fully AI-equipped car, but these systems are of course reliant on your phone and therefore suffer certain disadvantages. You might forget the device, or it could be out of battery. If your phone’s personal assistant technology is outdated, you’ll miss out on features.

Integrating this type of functionality in the cabin is what Nuance Communications is doing with its Dragon Drive technology. The system, which you can currently test in a special Chrysler Pacifica minivan equipped with two listening mics for the cabin, is a good poster child for what we can expect from early-generation in-car AI.

While voice-commands for your infotainment and climate control systems are not brand-new concepts, Dragon Drive adds a layer of functionality thanks to its Contextual Reasoning Framework:

  • The system remembers your preferences, and can, for example, display a favorite parking garage first rather than showing a complete list of parking options.
  • It can even make sense of another passenger’s voice and adjust the climate in their zone only.
  • The bleeding edge of the drive technology includes features like “just talk,” which eliminates the need for the customary “Hello Dragon” prompt that is Nuance’s equivalent to “Hey Siri” or “Ok Google.”
  • When Just Talk is engaged, the driver might simply mention that it’s too warm in the car, and Dragon Drive will intuitively bring the temperature down.

It remains to be seen how Nuance’s tech will compete in a marketplace where leaders like Apple and Google already have a claim, but it’s hopeful there are ways for the technologies to interoperate.

Ford Teams up to Offer Automation

If you’ve followed some of the major announcements at the North American International Auto Show, you know Ford Motors has big plans for the future. It starts with little things, like helping you find a parking spot, but Ford wants to offer a self-driving vehicle by 2021, along with potential additional transport services. It’s a holistic transformation from car company to mobility company.

Part of this plan involves the addition of AI features as standard with Ford’s in-car technology suite. To speed this process along, it’s partnered with Argo AI, a startup founded by former Google and Uber leaders specifically to develop automated cars.

While development of Ford’s autonomous vehicle will be the top priority for Argo over the next four years, there is still the potential to license to other companies down the road. In this way, both companies benefit. With so many tech leaders invested in delivering AI to the automotive world, it’s a case where the consumer wins thanks to healthy competition.

The Risks of Automation

With more automation in cars comes more attack surfaces for cybercriminals to exploit (Chesky/Shutterstock)

Along with the unavoidable question of just how well self-driving cars will drive, there are inherent risks involved with introducing network-reliant technology in cars. Throughout the 21st century, we’ve watched cybercrime become a more and more real part of our daily lives, and when cars become potential targets, there are bound to be some precarious situations.

Some type of increased security will have to be implemented, since the systems in place today are said to have more than 50 attack surfaces for criminals to target. In one recent example of car hacking, a team from Texas took over virtually every control surface in a Jeep Cherokee. Chrysler has since fixed the issue, and says its vehicles could only be compromised by someone with physical access to the car.

Researchers are already focusing their efforts on helping AI systems identify inputs from unique users. The challenge is in delivering the crisp response we expect from the controls in our cars. It’s not acceptable to have your car delay a brake input by a second to process whether it came from you. This is the dilemma that makes in-car security so difficult with AI.

A Paradigm Shift for Commuters

Five years is a relatively short amount of time for such a drastic change, particularly if we’re seeing completely self-driving automobiles by then. It begs the question, how will this change to one of our most basic infrastructure components and affect daily life?

Many have speculated we can look forward to increased productivity in the long term as commuters gain the ability to complete business tasks during the time they would typically be driving.

There’s also expected to be a renewal in the need for automotive technicians as the trade becomes more a matter of computer science than turning wrenches. An entire micro-economy is expected to spring up to provide services in-car while waiting to be chauffeured somewhere on autopilot.

A Time of Transition

It’s fine for automakers to promise cars that will drive themselves in the near future, but delivering them is very different. Two things seem certain, though:

  • We’ve already started to enjoy some of the wins modern AI can deliver in cars on the market today
  • The transition to full automation will be a process, not an overnight shift

Just as Tesla doesn’t endorse drivers taking both hands off the wheel while using autopilot, the first self-driving cars will be designed with fail-safe systems allowing the driver to step in to avoid an emergency. It is conceivable at some point in the future, we’ll have cars that are devoid of human-usable controls altogether, but to guess at whether that will take place in our lifetimes is impossible for now.

About the author: Kayla Matthews writes about artificial intelligence, big data and the cloud for websites like CloudTweaks, VentureBeat and Motherboard. To read more posts from Kayla, subscribe to her newsletter on Productivity Bytes.

Related Items:

How Auto Insurers Detect and Use Your Driving ‘Fingerprint’

Automakers Bet Big on Data and Analytics

The Big Data Behind Shell’s Super-Efficient Car Races