The Danger of Voice Commands and Driving: What Apple Siri, Google Now, and Car Companies are Missing
It’s 6:00 am. My alarm goes off, and while I wish I had time to watch the news, catch up on email, and check what my friends are up to on Facebook, I desperately need to get a workout in and drive to work in time to prep for my 8am meeting. I go for a run and hop in my car. I’m not a fan of texting and driving, but while crawling at 10 mph in traffic I innocently peek down at my phone to skim my unread email subject lines. In the time I glance down and look back up I need to slam on my brakes– traffic had come to a stop in the 2 seconds I looked away. Close call.
Sound familiar? It should. The National Highway Traffic Safety Administration estimates that at any given daylight moment, 660,000 drivers handle smartphones and manipulate electronic devices. Sometimes our hearts race when a car brakes suddenly. Other times it doesn’t end so well: the National Safety Council reported that motor-vehicle deaths increased to 18,600 in first 6 months of 2015, up 14% over 2014; the uptick partly attributed to driver distraction due to phones.
We’ve all heard campaigns that tell us to stop texting and driving. But the fact is that our dependence on smartphones is utterly complete. We feel uncomfortable when we can’t access them.
Google, Apple, and car companies have all created voice command systems to help address this issue, but it’s not working. In fact, a 2015 JD Power Vehicle Dependability study reported that most frequent car complaints are now related to Voice Command systems. Even worse, recent research by AAA Foundation of Safety seems to suggest that Voice Command systems are often worse than texting and driving.
So, what’s missing?
The big gap: browsing vs. analytical voice commands
What makes voice commands difficult to use in the car is not the unreliability of the interface in noisy environments, the fact that it is mentally tasking, or that it takes seconds or more for each interaction — though those contribute of course — but that conversational interfaces, even when perfect, do not address the way we want to process information. When we are stuck in traffic jams and bored behind the wheel, we aren’t looking to yell commands at an analytical conversational interface. We are looking to browse the web and pick out what we feel like paying attention to. Like a kid in an ice cream shop, we often do not know what we want until we browse.
Browsing is different from the analytical strategies we use with voice command systems such as such as Siri & Google Voice. What is the difference?
Browsing: This is the most common way you use your phone in a given day. You simply scroll through apps or websites of interest (twitter, email, reddit etc.) casually taking in information rather than zeroing in on a specific question that you want answered. It is not mentally taxing and can be done for hours at a time - as you are probably already aware.
Analytical strategies: Much more akin to how you search through an encyclopedia for an answer to a specific question, this is the current method of search used by Siri or Amazon Echo. You proactively hunt for the answer to a single question at a time. The search terms used are incredibly important and, if imperfect (as they often are) lead to incorrect answers, or no answers at all.
In short: one is effortless and relaxing, while the other takes work. No matter how “perfect” a voice command interface is, it doesn’t address the core differences between these two methods of interaction. At FingerTips Lab, we are building a platform that lets you check your email, twitter and browse your newsfeed without ever speaking a word or taking your eyes off the road. Consider this the most personal radio station you’ve ever listened to.
Interested in learning more? Sign up at www.o6app.com.