When I was growing up, talking cars were the stuff of science fiction. On the show Knight Rider, Michael Knight worked alongside his car, KITT, to solve crimes across the highways and byways of America. Long before Knight Rider, we had Dick Tracy. His two-way radio wristwatch might just be on your arm today, paired to your smartphone.
In recent years, the notion of a talking car has become something of a reality. With in-car infotainment advances from some of technology’s biggest players, drivers can use their voice to ask their car to play music, call their mothers, and even route them toward the nearest cheeseburger stand.
In-car voice control, however, is still in its infancy. Voice assistants will take commands, but they don’t converse. When you ask it to play a song, it can guess which artist you want, but it doesn’t ask for clarification; and because of the third-party nature of most in-car AI today, voice controls tend to be tied to what is happening on your smartphone—not what is happening in your car.
The future of AI-enabled cars
From my perspective, the future of true AI-powered comfort and convenience for drivers isn’t far off. Thanks to Natural Language Processing (NLP)—a technology I've been working within my business, Passage AI—automakers have the potential to build AI-powered chatbots capable of creating customized, cross-device experiences. These chatbots could answer common auto-intender and car owner questions, and make life safer and easier for people behind the wheel.
At Passage AI, we're currently working with several automakers to develop different use cases, including
programming a spoken owner’s manual. We're also working with Voice to enable functionality such as opening the hood and opening and locking doors. Anything that an automaker assigns to an app can be turned into a voice function.
Imagine a world where your car asks your preferred temperature, seat height and mirror positioning, and adjusts to all these the moment it hears your voice. Or when the oil light falls below a certain level, your car asks you where and when you’d like to book an oil change—then books it for you and adds it to your calendar. If you get a flat tire, would you like roadside assistance called, or would you prefer step-by-step instructions on how to
change the tire yourself?
While smartphones are powering today’s voice communication in cars, I believe that we’ll see the car-based AI of the near future turn this paradigm on its head, extending interactions outside of the vehicle to phones and other smart devices.
The challenge for car-native voice assistants
Thanks to the proliferation of digital home assistants, automakers plan on extending the trend to cars. But there are potential hurdles in making the leap from home to car, such as getting voice commands correct in every context. Making an error with a voice command in the home can have annoying or humorous consequences, but making an error on the road can have disastrous consequences.
Each voice assistant must understand a wide array of dialects, accents, idioms and intonations. The in-car assistant also must understand and separate background noise—the engine, windshield wipers, other cars passing, music and children in the back seats—from the driver’s voice. Setting up controls on where the voice comes from and to ignore other passengers is one way to mitigate the challenge. Ultimately, evaluating the efficacy of these in-vehicle voice systems in various real-world conditions is critical.
While today’s smartphone-powered, in-car assistants provide an adequate solution for guidance and entertainment, I believe they’re ultimately a stop-gap solution on the path to better technology delivered by the automakers themselves. What’s more, the quality of these offerings may even become a determining factor in the car-buying experience. It’s not uncommon for a salesperson today to hear questions like, “But does it support Android Auto?”
What happens now?
With manufacturer solutions giving better, more car-oriented experiences, personal AI is not too far out of reach for the automotive industry. Google and Amazon are working with various carmakers to integrate their
respective voice assistants into future models. Mercedes-Benz has already introduced its in-vehicle voice assistant, enabling drivers to talk directly to their cars and have the cars talk back. BMW and other automakers are quickly revealing their plans to capitalize on the voice assistant trend. These systems are poised to improve over time, learning and adapting to the needs and preferences of the driver.
From my experiences working with these technologies, I can say that conversational artificial intelligence and NLP will be a game-changer in the driver experience, and potentially a future standard in automobiles just like smartphone-powered voice control options have become today.
Regardless of what the industry standard becomes, though, I believe any automaker that takes the initiative of creating a custom experience tailored to its vehicles can put itself in a winning position when it comes to creating great driver experiences.
Mitul Tiwari will be sharing the stage with fellow experts in the auto and voice-first field during the session, "Designing Next-Level In-Car Experience," at VOICE Summit on July 24th. Register here to attend this and dozens more sessions!
*This guest post has been lightly edited for clarity. Submit your post here to see your own ideas on the VOICE Blog.