Posted on 22 March 2013.
Photo: Alex Washburn/Wired
Apple has a vision of a future in which the disembodied voice of Siri is your constant companion.
It goes something like this: You arrive home at the end of a long day and plop down on the couch. A beer in one hand, your phone in the other, you say, “Siri, open Netflix and play The IT Crowd.” Midway through the program, you feel a draft. “Siri, it’s cold in here.” Siri politely tells you the temperature, and asks if you’d like it raised. The furnace kicks on. As the credits roll down the TV screen, Siri reminds you of your dinner date downtown. In the car, she gives you turn-by-turn directions to the restaurant and sends your date a text message to say you’re on the way. Halfway to dinner, you realize you need movie tickets. No problem. Siri takes care of that, too.
This is where Apple is headed with Siri, as the nascent voice-activated AI spreads from our phones to our desktops, our homes and even our dashboards to become our concierge to the digital world. Cupertino is moving aggressively to develop a distinct personality for Siri that will make interaction more natural and fluid, and kindle the innate human tendency to anthropomorphize objects.
So far, Apple’s results have been a mixed bag at best. But its ambition is breathtaking, and if the company succeeds, Siri will ultimately revolutionize how we interact with our phones and our environment.
“We spend so much time with our cellphones that having an effective personal assistant could be revolutionary,” said Andrew Ng, director of Stanford University’s AI Lab. “Most people spend more hours a day with their cellphone than with their spouse. The cellphone has the potential to learn your personal habits and preferences, and serve you in a very personalized way.”
To do this, Apple must catch up with, and then overtake, Google, which offers voice search capabilities and natural language understanding far superior to Apple’s. Google also has a technical advantage, with its deep back-end engineering approach to product development, while Apple is known more for design and UI. But as Siri’s technical abilities improve, Apple could become the first company to introduce a truly useful speech-recognition tool with a humanized AI that doesn’t make you cringe or throw up your arms in frustration.
Siri debuted to much aplomb with the launch of iOS 5 and the iPhone 4S in 2011. But her limited capabilities and service outages that rendered her speechless showed just how beta she was. A handful of OS updates have expanded her original feature set and capabilities, but in many ways Siri remains a broken promise.
But there are signs of progress.
With iOS 6, Siri gained several features, including the ability to sift through sports stats, mine movie data from Rotten Tomatoes and make reservations through OpenTable. The latest major update, iOS 6.1, lets youorder movie tickets through Fandango. Of course, she can coordinate everything through your calendar, messaging and e-mail apps. She does some of these things better than others, and she still has difficulty recognizing a variety of verbal queries that make those tasks useful.
But recent job listings suggest Siri is moving beyond her current skill set in a big way.
“Do you want to be a part of the next big revolution in human-computer interaction? Are you looking to contribute to a product that is redefining the smartphone?” a January posting for a Siri software engineer asks. The engineer will explore “new areas of expertise for Siri, expanding the product’s capabilities for millions of users.” Another Siri engineering position will fill a vacancy on a team looking at Siri as “an entire miniature OS within the OS.”
Clearly Apple wants Siri to do things iOS or Mac OS X can do, like multitasking and working with third-party apps. That suggests Siri will become a unique ecosystem.
The first step must be a public Siri API. Building out a robust API for third-party developers could do for Siri what the App Store did for iOS — make it a rousing success. Developers are eager to hook into Siri to increase engagement and make interactions more natural and fluid.
“When we saw Siri, we saw it as an evolution of an assistant and personality,” Mark Young, vice president of mobile strategy at Fandango, said. “For us it was all about working with Apple to compress users’ intent to action — getting them to buy a ticket in the area more quickly. Siri lets us seamlessly link discovery to buying to redemption.”
Other developers see great things for Siri.
“We’d love to see Siri and Hipmunk integration for questions like, ‘Is there a hotel with availability near me?’ or ‘Which hotels are in the foodie neighborhood?’ Those are quick questions that can use quick answers, and I think Siri could help quite a bit with that,” Hipmunk CEO Adam Goldstein told Wired.
Introducing Siri to the app ecosystem brings greater convenience to consumers. By disambiguating information found in an app, you don’t need to swipe and tap your way into an app interface. The idea of a closed off, insular app fades away. You can use Siri to open the app, get information, make purchases. Your only reason to enter an app would be to play a game or watch a video.
“There is no need for me to open my calendar and look at the day when Siri can tell me,” said Ben Bajarin, director of consumer technologies at market research firm Creative Strategies. “Or with my e-mail, Siri can tell me about new e-mails or read them in the future. And while using Siri for something related to that application, I don’t necessarily need to open it.”
But if Apple succeeds in its plans, Siri won’t be limited to your smartphone. She’ll be everywhere you are — in the office, at home and on the road.
Automakers already love Siri — it’s one reason Apple executive Eddy Cue got a seat on Ferrari’s board of directors. Rather than developing expensive proprietary speech-recognition systems, a dozen car manufacturers, including Chevrolet, Honda and BMW, are leveraging Siri to deliver hands-free controlof automotive infotainment systems. The Chevrolet Spark and Sonic were the first to incorporate “Siri Eyes Free.”
“You get the things you want from Siri in your car, hands free, your eyes on the road, while not being tempted to look at or touch your phone while you’re driving,” GM spokesman Scott Fosgard said.
Before long, Siri’s voice interactions in the car will be more anticipatory. Let’s say you’re stuck in traffic and will be late to that appointment you’ve got in 15 minutes. Ideally, Siri might notice your dilemma, estimate your travel time and send your client a message with an update — then find the nearest available parking space once you’ve arrived.
You can also expect Siri to slink onto your desktop sometime soon, at least in some form. Apple already has speech recognition capability in OS X, but it makes sense to bring Siri into the fold.
Bajarin thinks on the desktop, Siri will be especially useful when it comes to search. “With Siri, Apple has a way to provide a value layer on top of search and make it even more relevant,” Bajarin said. “Apple will continue to use Google for search, but Siri lets them add an element of control on top of Google that can be integrated nicely into Apple products.”
Siri’s most exciting role, however, could be managing our connected homes. With your iPhone or iPad as a central hub, you might be able to use speech commands with Siri to control everything from the lighting to room temperature, or start warming up the oven from the couch. Siri will make channel surfing and movie watching a piece of cake as your phone, tablet, TV or set-top box respond to your spoken requests and queries.
Expanding Siri’s capabilities is but one part of Siri’s grand future. Another recent Apple job listing suggests the company is adding character-driven dialog to give Siri a distinct personality — moving into the UI minefield of anthropomorphic technology.
Past attempts to make technology human-like have largely failed — remember Jeeves? How about Clippy? But done right, slapping a personality on Siri could make her — or him; why not? — more appealing as an ever-present companion somewhere between a butler and a beloved pet. People, after all, name their Roombas and cry during Wall-E.
“Siri in its current state is more of a utility rather that something that is emotionally attractive,” said Thilo Koslowski, a technology analyst at Gartner. “I think adding the personality and the character to Siri, and having people choose those characters, would be something sticky for consumers to choose to use.”
It’s a tricky line to walk, as Microsoft can tell you. Clippy was a disaster, and did anyone like Power Pup? The difference here, Forrester analyst Charles Golvin said, is Siri responds to requests; she doesn’t interrupt. “That’s an opportunity for annoyance that’s not there.” he said. “Siri is not proactive.”
Like with every other industry Apple has come to revolutionize, Apple has watched other AI’s and humanized helpers come and go and learned from their failures. Siri is evolving slowly, clunkily. Whether by design or accident, that slow maturation is giving us time to grow comfortable with her expanding capabilities and developing personality.
But there are still problems to overcome. While vast improvements have been made in recent years, speech and voice recognition aren’t yet perfected. Siri’s chief competitor, Google Voice Search, offers far better speech recognition, particularly in the car, where ambient noise and echoes can strain the technology.
Google also offers superior natural language understanding, not to mention better recognition of dialects and slang, which allows for more natural interactions.
These are the engineering challenges facing Apple, which, traditionally, hasn’t had the depth of engineering Google has. Google’s rich background in search, particularly regarding its robustknowledge graph, which identifies context in what you’re typing or saying in a search query, is an especial boon to the back-end of its voice search function. Apple now has to build up its own search and contextualization algorithms so that Siri can not only understand what you’re saying, but what you actually mean.
Over the next few years, Apple will continue tracking when and how we use Siri, when she fulfills our needs, and when she fails. This information will be used to steadily improve and expand her capabilities as she evolves from a mediocre disembodied note taker to a full fledged, humanized AI assistant.
There’s a tremendous opportunity: to introduce the first virtual helper that is both a utility and a personality, one that can infiltrate nearly every facet of our lives. If it succeeds, it will change everything.