adimas - Fotolia
Auto manufacturers are facing sales headwinds as fewer millennials buy cars. And with the future of driving looking more and more autonomous, the industry is in for even greater upheaval ahead.
But audio and connected device maker Harman International Industries Inc. is betting that tapping into more consumer technology, particularly by putting AI in cars, will play a big role in smoothing manufacturers' road into the future.
"[Vehicle manufacturers] are worried that a lot of millennials aren't buying vehicles," said Stephen Surhigh, vice president and general manager of cloud services at Harman. "So to appeal to some of the tech-savvy buyers, to incorporate the features that they know and love from their mobile device, it's all part of what automotive manufacturers have to do today."
Virtual assistants that sound like your brand
Harman is working with manufacturers to build more connected technology into vehicles.
At last month's IBM Think conference, the company demonstrated a virtual assistant tool built using IBM's new Watson Assistant technology that was embedded in a Maserati. Surhigh said the idea is that AI in cars will be able to proactively read your calendar, pull up directions to your next meeting and build in a stop along the way if you need gas, all automatically.
This is one of the first implementations of IBM's Watson Assistant, which is intended to enable enterprises to build their own AI virtual assistants similar to commercially available versions like Alexa, Google Assistant or Siri.
Watson Assistant uses machine learning to remember user preferences and natural language to interact with users. A business can customize its voice and persona to fit its brand. This means that every implementation could, theoretically, be unique rather than simply featuring a Watson persona.
Car companies need to get with the times
However, the Harmon technology demonstrated in the Maserati at IBM Think isn't available to the average consumer just yet. Surhigh said part of the problem is that car manufacturers have three-year development cycles. So technology that becomes available today might take three years to find its way into cars out in the wild. That is too long, according to Surhigh.
"Waiting until 2021, you're missing multiple consumer electronic cycles," he said. "We have to get better where it's not a three-year cycle."
Harmon is working with manufacturers to speed up this process. Part of the answer might be building more powerful processors into vehicles. Surhigh said today's vehicles ship with relatively small processors that are essentially maxed out with existing demands. This means that manufacturers can't simply send a software update the way a mobile device company might. It would take a more substantial processor to run AI in cars.
Implementing this kind of technology is only going to become more important as the industry moves into the autonomous car future. Surhigh said it's hard to know how things will change exactly, but ownership models might shift. Instead of every family owning a car or two, they might subscribe to some kind of sharing service where they order a car as needed.
Either way, Surhigh said brands will need to offer a unique experience to attract customers, and the kind of personalization offered by AI in cars will play an important role in this. "It's about making the user experience in the vehicle more seamless," he said.