I’ve spent the last week talking to my phone. Not that revolutionary you might think; it is after all a phone, and voice has been part of the feature set right from the get-go. But in this case I’ve been testing out Siri, one of Apple’s key selling points for the iPhone 4S. Siri allows you to ‘talk’ to the phone in order to make calls, appointments, send messages and search for information.
In one sense this is nothing new; older smartphones, including those from other platforms have had voice control features for many years now. Where Siri makes it interesting is in its ability to handle natural language. Where most of the other systems rely on very simple phrases, Siri can handle longer contextual strings and a variety of voice inputs. So you could say, for example, “What’s the weather like in Melbourne”, and it’ll find a five day forecast; ask it then “what’s the time there?” and it’ll remember the context and give you AEDST time for Melbourne. It’s all rather reminiscent of Star Trek, frankly, speaking to a small computer in your pocket, although it does rely on having a net connection of some sort as some of the voice processing is done at Apple’s servers rather than in the device itself. It learns as it goes, according to Apple, so rather like products such as Nuance’s Dragon Naturally Speaking, the more you use it the better it’ll get. It’s also got a specific setting for English (Australian), and it’s highly advised that you use it; the difference in its understanding of a strine accent and a yankee one is remarkable. There’s obvious scope here for use by those with physical ability limitations where typing is difficult or impossible, but even just as a cool gimmick.
Siri does have its limitations, especially locally. Ask it for any kind of directions, and it’ll sadly inform you it can only give directions when it’s in the US. That’s not reticence on its part; the directions part of Siri’s logic relies on a couple of US-specific services that Apple’s signed up to. There are words and phrases it’ll stumble on repeatedly, and because it learns its owner’s voice, it’s markedly less effective for other users if they borrow your phone. To be fair to Apple, it does mark Siri as a “beta” (that is, still in development) product, so there’s plenty of room for improvement.
There’s arguably a bigger strike against Siri to consider, however, and it’s true for any voice controlled product. Within the context of your own home or office, talking to a computer may feel a little odd at first, but generally you’ll have access to a keyboard anyway, which adds a layer of instant precision. Out and about, and you’re going to have to talk over the general chatter and noise of the world, which means relatively loud. All of a sudden, you’re getting Siri to calculate the interest on your home loan, or noting the times of your medical appointments in public. Most of us would rather keep that stuff private. There’s no easy way around that — it’s decidedly a public perception problem rather than a technology one.