Friday, October 14, 2011

Will Apple's 'Siri' Bring Voice-Control Into the Mainstream?

Many critics were disappointed at the lack of any iPhone 5 announcement from Apple, and whilst there is no flashy new design for the exterior, the iPhone 4S has a number of major new features. One of the main features being touted was Siri, a voice-controlled software assistant. Siri allows you to interact with your phone using just your voice. It can set calendar events, even checking for schedule conflicts, it can check weather and stock reports and find answers to questions using the powerful computational knowledge engine Wolfram Alpha. This seems like an impressive offering, but despite improvements in the technology, speech-recognition has not found it’s way into mainstream products so far, can Apple be the first to achieve widespread adoption?

Speech recognition is already well established. Most people will have encountered, and probably been infuriated by, voice recognition in automated telephone support systems. But voice recognition is also an important assistive technology. Both Microsoft Windows and Mac OSX operating systems have extensive built in voice control and screen-reading systems to allow all functions to be automated via voice commands.

Nuance’s speech recognition software ‘Dragon Naturally Speaking’ is the main choice, and has advanced dramatically over the last decade. Several years ago, their dictation software required a lengthy calibration session of an hour or so to allow the system to learn the user’s voice; today, you can use their free mobile app to dictate text messages and activate functions with no calibration required whatsoever.


 But figuring out the actual words that have been spoken is only the first part of the challenge. The spoken words must then be interpreted into some kind of query or command that the computer can understand and then act on. Making sense of natural language, with all it’s ambiguities and turns of phrase, has been an incredibly difficult challenge so far.

Recently, IBM’s ‘Watson’ computer software managed to win the popular game-show ‘Jeapardy’, playing against previous show champions. The game show was chosen because not only does it require a great breadth of knowledge, but the questions are often posed using ambiguous or cryptic language. The Jeapardy challenge was a great PR stunt for demonstrating the software, but this marked a significant leap forward, opening the door to natural language communication with machines.

Apple has a long-standing tradition of bringing products and technologies into the mainstream and achieving widespread adoption: the iPod was by no means the first MP3 player on the market, but it’s success quickly made the label ‘iPod’ into a generic term for portable MP3 player; the iPhone was the first touch-screen only device; and the iPad still remains by far the most popular and sought-after tablet despite numerous aspiring competitors. This is not only down to clever marketing, Apple has achieved widespread consumer adoption of their devices for their ‘it just works’ approach to ease of use. Customers trust that not only will their products be easy to use, but they will also work well together.

So can Apple turn Siri into the first major mainstream voice-controlled application? People will have to get used to interacting with their mobile devices in a new way, but if Apple can achieve widespread adoption through it’s products, then this will no doubt drive further innovation of the technology and will see it’s implementation in many more products and applications all around us, and we can all look forward to fascinating conversations with our washing machines and public lifts.

1 comment:

  1. that's fabulous book of the Software Society.Apple Siri is so good.


Note: Only a member of this blog may post a comment.