Several news items from the Mobile World Congress make it clear that Siri – the native, speech-enabled mobile assistant running on the iPhone 4S – is the benchmark against which competing mobile speech applications wish to be measured. It’s very good news for Apple (a company whose $500+ billion market cap exceeds the GDPs of Poland, Belgium, Sweden, Saudi Arabia, and Taiwan.

It started with a series of video press releases from Google’s soon-to-be device manufacturing subsidiary, Motorola Mobility. Under the headline, “Google Voice Actions Knocks Out Siri on Motorola Android Phones”, readers could see how Google Voice Actions on several models of Moto’s wireless phones carried out dictation, getting directions, posting a reminder and Web navigation tasks faster than the iPhone 4S.

One of the “knockouts” is rendered below:

I first wrote about Voice Actions for Android in August 2010 noting that “When we look back on the summer of 2010, the launch of Voice Actions for Android will be seen as a signal event. It goes a long way toward re-establishing the spoken word as the natural input for a phone (duh!).” That may have been true and it certainly rattled the chains of Microsoft (which spun off Tellme, but kept much of the IP that will power voice and gesture controlled devices) and Nuance (which, in addition to carrying out the automated speech processing for Siri, is in the process of acquiring Vlingo and continues to refine Dragon Go! and a number of initiatives that define a more Natural User Interface).

What the boxing match between Voice Actions for Android and Siri fails to capture is how real people actually use Siri. This post provides details about one person’s daily experience. While the notorious “sample of one” is indicative of only one subscriber’s preferences, it is fascinating to see how usage starts in the morning with a query about the weather and crescendos into classic Q&A exercises, giving a person a verbal way to ask someone a question before it recedes in memory. That’s the real-life use case (mostly because the idea of reciting text messages aloud fundamentally misunderstands the purpose of “passing electronic notes” in private and asking for directions aloud when you still have to touch the screen and look at a map defeats the purpose of “hands-free/eyes forward” navigation.)

But I have to say that this announcement, coming out of MWC, caught my imagination. A joint venture between embedded ASR specialist Novauris and Existor Ltd, a technology firm that applies artificial intelligence to the challenge of creating “conversational,” spoken interfaces. The two companies plan “to combine their expertise in artificial intelligence (AI), natural language processing and automatic speech recognition (ASR) to create mobile and wireless applications that allow people to interact conversationally – not just with smartphones and computers, but also with cars, refrigerators, TVs, toys or any device equipped with a suitable embedded processor.”

The ambition to make consumer electronic devices, appliances and automobiles more conversational is very compelling. Nonetheless, I believe the most interesting aspect of the Novauris/Existor output is the potential to remain conversational even in a disconnected state. Many Siri users have told me that their mobile assistant will duck a question now and then for no apparent reason. They speculate that access to the server “in the cloud” may have timed out. At that point Siri will ask for the user to try later or default to a Web search. In their announcement the two technology companies claim that their first output will be a series of speech-enabeled “chatbot apps,” called CleverApps which will run on smartphones without any need for access to “the cloud” over the data link.

In addition to solving the problem of getting an answer in a disconnected state, the new venture’s apps have the potential to reduce data traffic at a time when some wireless carriers are “throttling back” the speeds for heavy users and getting rid of “unlimited data plans.” With product announcements and vendor contests coming fast-and-furious, the ripple effect of Apple’s Siri launch is gathering momentum.

Categories: Articles

Tags: Nuance, Apple iPhone, mobile speech, Siri, Speech Actions for Android, AI