Kitt, the famous talking car from the TV series Knight Rider, is being brought to “life” by Robin Labs, an Israeli-California start-up that has designed an intelligent voice recognition and communication system for Android devices.
Robin’s technology is designed for to get drivers updates on traffic, routes, parking locations, gas prices, weather, and more. It can even, Company co-founder and CEO Ilya Eckstein says it can even tell jokes.
In that sense, said Eckstein, Robin’s technology, now featured in an Android app but ready to be embedded in a host of add-ons and devices for cars, is a lot like the science fiction intelligent robotic “characters” many of us grew up on — from the Enterprise’s service-oriented computer that kept the various Star Trek crews alive in deep space, to the malevolent craftiness of 2001’s HAL, to the corny jokiness of Kitt.
To some, those characters were science fiction novelties, but to others, the idea of an intelligent computer that could understand what it was asked — in context — and respond appropriately was a dream to be followed. Eckstein told The Times of Israel, that today, the dream is within reach. “We have the location technology, the fast database searches, and the wisdom of crowds to support services like Waze, Yelp, and others that bring information to users,” he said.
The issue of getting a computer to understand language and its nuances is another matter altogether. “There’s a popular 80s song titled ‘Call Me,’” said Eckstein. “Let’s say you wanted to hear that song on your car music system. If things worked ideally, you should be able to say ‘Call Me,’ and the system will play that song,” he said.
“But what if the system also lets you make phone calls, and lets you use your voice to dial contacts. How is the system supposed to know that ‘Call Me’ means the song, and not ‘call a contact by the name of Me from my contact list?’ Another good example is “Boston,” said Eckstein. “Are we talking about the city, the music group, or the fast-food restaurant Boston Market?”
That’s exactly the problem with Siri, the not-so-advanced voice technology by Apple, said Eckstein. “If you tell Siri ‘call me an ambulance,’ then that is exactly what Siri will do — address you with the term ‘ambulance.’ That’s not natural language communication, not the way we at Robin Labs do it, at least.”
Natural language understanding (NLU) technology for artificial intelligence applications like the ones produced by Robin Labs has been used for several decades, but one thing has been missing, said Eckstein — context. That would enable the computer to figure out which out of a variety of possible responses is the right one. “Once you get over that problem, you can build many apps that intelligently use conversation, with inputs by users and responses by the machine.”
Fortunately for Eckstein, there are many places to draw on for context these days. By searching through data sources, like a user’s music library and calendar, crowdsourced conversations on Twitter and other social networks, and examining a user’s history, Robin’s system can make much better and more intelligent “guesses” about what it’s being asked, and how it should respond. “People talk and interact, and we gather that information and apply it,” said Eckstein. “The result is a much more accurate conversation exchange.” It can tell jokes — and even respond to them, since it could probably figure out the punchlines based on social media conversation, the company said.
The Robin app is available free for Android devices, but an iPhone version is not in the works for now. It offers features such as local search (including Yelp reviews), navigation, real-time traffic and parking information, fuel prices, weather and more. And for those getting bored behind the wheel, the company said, Robin can narrate personal Twitter news, with more content reportedly coming soon, and of course, tell jokes. Like Siri, though, not all features, such as gas prices and Yelp data, will work in all geographic areas.
Although the Robin technology could apply to a plethora of use cases (imagine a cooking app that lets the cook ask questions about adjusting ingredients, substitutions, etc.), Robin is starting with cars, developing apps and devices to help drivers. The automobile industry is notoriously slow to adopt new technology, especially in a vehicle’s entertainment component — so, said Eckstein, it’s likely that Robin will deploy its technology in add-on devices for cars.
At CES in January this year, Robin, together with Pioneer, well known for its car entertainment systems, introduced a voice-driven mirror that will act as a navigation screen, messaging center, and entertainment interface. The companies are working together to perfect it before putting it on the market.
“Today, people still think of a car assistant primarily in terms of navigation,” said Eckstein. “But in fact, we don’t use navigation much, as most of the time we know our way around. What we really need in the car is someone to watch out for us as we go and be there when we need… well, just about anything. That, and being able to remain fully connected and empowered behind the wheel. To not miss out on a single important bit of life, even when you are not staring at the screen. And that is exactly what Robin is meant to help us with.”