Skip to Content

Cars May Soon Understand More of What You Say

It should soon be possible to give your car more complicated and natural verbal commands.
July 21, 2015

Many cars now come with voice control, but you can’t really talk normally to such systems, and you often have to repeat a phrase to get the job done. That could change, however, with the introduction of voice interfaces that allow for a more natural back-and-forth between driver and dashboard.

“What we’re going to see in the very near future is the ability to have a dialogue,” says Charlie Ortiz, who is senior principal manager of the artificial intelligence and reasoning group at Nuance, a voice recognition technology company based in Burlington, Massachusetts. “You might say I want to listen to some Latin jazz, or suggest a particular musician.”

Ortiz says that such technology is now in the vehicle production pipeline, which means it may appear within a few years. It will primarily allow for more natural control of dashboard features and retrieval of information such as directions. “In the navigation domain, we’re developing methods to describe points of interest more abstractly,” he says. “I don’t always know the exact address of where I want to go. I want to be able to say ‘I want to go to a restaurant in the marina near the ballpark.’ “

Nuance came to dominate the market for voice-recognition technology over the past decade after acquiring various other companies in that space (see “Where Speech Recognition Is Going”). Thanks to new techniques and large quantities of training data, speech recognition has improved greatly over that time, and Nuance supplies the technology to companies across numerous industries. It already provides voice control technology to carmakers including Ford, Hyundai, and Chrysler.

Nuance is now looking to build on that by offering greater understanding of speech. This is notoriously difficult, though, because the meaning of words and sentences can vary dramatically depending on the context; and so dialogue usually needs to be carefully constrained within certain areas. Conducting more complex conversations is a major goal for the lab Ortiz runs at Nuance. His team is working to develop personal assistants capable of understanding more types of sentences and responding effectively when they do not comprehend. For instance, a query might refer to a previous discussion and require a subtle appreciation of its context. A user might ask such a system how a particular restaurant compares, in terms of user rating, to other restaurants in his or her search history.

And Ortiz believes that more fluent speech technology could be just around the corner, thanks to advances in parsing semantics. “The stars are aligning at just the right time,” he says. “There have been a lot of advances in various components—language-understanding and the reasoning back-end parts. One big challenge is to put these pieces together.”

Another key challenge, as far as the auto industry is concerned, is ensuring that more sophisticated interfaces aren’t also more distracting. More intuitive speech interfaces might be less taxing, but only if they work well.

“If it works perfectly, great. If it fails, you’re in a worse position,” says Bryan Reimer, a scientist at MIT’s Age Lab, whose research has shown that voice interfaces can be just as distracting as regular ones in cars. “The more complex and vague the commands, the more complex the recognition problem, and the higher damage of failure.”

Several carmakers contacted by MIT Technology Review declined to discuss how voice technology would likely evolve in their products. However, vehicle interfaces are advancing at an impressive pace, spurred on in part by mobile technology (see “Rebooting the Automobile”).

Keep Reading

Most Popular

Scientists are finding signals of long covid in blood. They could lead to new treatments.

Faults in a certain part of the immune system might be at the root of some long covid cases, new research suggests.

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.