Apple is bringing speech recognition to the masses with its new iPhone 4S, equipped with an intelligent assistant named Siri. It’s a major differentiator for the new iPhone, setting it apart from its predecessors. I’ve been using speech recognition software for the past 8 years, so I was eager to take this enhanced version of Siri for a spin. Here’s my review.
So now Siri is baked into every iPhone 4S, and not available elsewhere. Siri has come a long way since it was first introduced as a less-accurate and somewhat incomplete iPhone app. Now it’s better integrated into iOS 5, and my immediate impression is that it’s more accurate than it’s ever been. Even in a noisy environment inside a car going 60 miles an hour, it can still understand most of what you’re saying if you hold the iPhone up to your ear. Its speech recognition isn’t perfect, and some of its errors are laughable, but in a quiet environment its accuracy is nearly equal to that of the desktop version of NaturallySpeaking running on extremely powerful processors.
Its integration into the iPhone 4S’s iOS 5 software makes it convenient to use. You press and hold the iPhone 4S’s Home button, and it springs to life, sounding a short beep to signal for you to begin speaking. You can use it in this speakerphone mode, or if the iPhone 4S is turned on, you can simply raise the handset to your ear (a necessity when riding in a noisy vehicle) and the phone’s proximity sensor activates Siri, usually prompting you to begin speaking (inexplicably, sometimes it doesn’t respond).
“Only a company with the chutzpah of Apple would have the courage to try something like this. But Siri works just barely well enough for Apple to pull it off, bolstered by the iPhone 4S’s faster processor and better camera.”
That odd non-working tendency must be why Apple is still calling Siri “beta.” The company reassures users that the Siri will be continuously improved, adding that the software learns how you speak as you go and will perform more accurate recognition as it learns your way of speaking. Still, loading beta software into a new piece of iPhone hardware is a thin thread on which to differentiate this new product. Only a company with the chutzpah of Apple would have the courage to try something like this. But Siri works just barely well enough for Apple to pull it off, bolstered by the iPhone 4S’s faster processor and better camera (among what Apple boasts as 197 other incremental improvements), all doing their part to strengthen the lure of this updated iPhone.
Over the 48 hours I’ve been using Siri, it’s hard to tell if it’s actually improving its speech recognition, but as it stands, it’s just good enough to be fun to use. I especially like the way you can almost carry on a conversation with it. For example, you can ask it, “How’s the weather in New York today?” It will answer by showing you the iPhone’s weather app with New York’s data displayed. Then, if you ask it, “Where are the good Italian restaurants there?,” Siri responds by finding 24 Italian restaurants in New York, sorted by rating. It knows you’re still talking about New York. Clever.
As you can read in our posts about Siri, it does bring a slight attitude along with it, which I find refreshing. Other times, it has hilarious misunderstandings, such as when I asked it yesterday to “Call me an ambulance,” and it responded, “From now on, I’ll call you ‘an ambulance’. Okay?” I was disappointed to hear Siri’s voice, which still sounds way too robotic for my taste. I was thinking that somehow, now that Apple owns the app, it would gussy it up to sound more like GPS units do, or like the mellifluous yet mutinous HAL 9000 from 2001: A Space Odyssey. But I suspect that’s still way off in the future. Instead, there are some oddities in its stilted pronunciation, such as the way Siri says the word “restaurant,” speaking with a drawl that sounds like it’s straight out of my native Southern U.S.: “Resta-runt.” Grandma, is that you?
Among its myriad capabilities, of course Siri can help you place phone calls with aplomb, where all you have to do is speak the name of anyone in your Contacts app, and it quickly connects you (something that’s been possible for years with much lesser cellphones). Beyond that, it can also help you speak an email and turn it into text, where it walks you through by asking who you’re sending it to, the subject line and so forth. However, it’s not too adept at breaking out separate paragraphs of text, even if I spoke to it the way I do with NaturallySpeaking, specifying things such as “new paragraph.” Although the email function could be useful for creating short emails while driving (not recommended), it still has some polishing to do before it’s truly useful for sending emails solely by speech.
Some of its capabilities go deeply into science fiction territory, such as pushing and holding the Home button, and then telling it to set a timer for 15 minutes. I especially like telling it to set an alarm, asking it directions, or asking it to launch a playlist in iTunes. I was disappointed to see that it wasn’t able to interact with Twitter, but I found a workaround for that, so that problem is solved already. Still, Apple should have made that capability available from the beginning, and if the company follows through on its promise, we will soon see a lot more interaction with various iOS apps.
Siri on the iPhone 4S still feels like a work in progress. I think it could have used another few months of development before it was released to push it well beyond gimmick territory. But Apple was already later than usual in its product cycle with this iPhone 4S, so might have been compelled to release it early. Even so, Siri as it stands now gives us a hint at what’s to come, and the future looks bright.