The limits of speech-recognition technology
For several years, speech-recognition technology has progressed rapidly, appearing in devices ranging from Bluetooth headsets, to phones, to cars. When first introduced, such devices required the user to speak slowly and deliberately, somewhat robotic-like. Recently, improvements to this technology have paved the way for natural speech-recognition, popularized by Apple when they introduced the ‘Siri’ feature on the iPhone 4S.
Having seen much success, Apple’s competitors such as Google and Microsoft are hard at work to renovate their voice-recognition capabilities to compete with these newer Apple iPhone models.
Natural speech-recognition, in theory, has the ability to decipher ordinary, naturally-spoken phrases to generate search results, connect to contacts, operate various applications, and more. Clearly, an updated feature like this would provide many advantages to users faced with a busy day and could (with the capability of hands-free, voice-operation) reduce phone-related accidents on our nation’s roads due to distracted driving. However, with the promise of such natural speech-recognition, one must wonder where the use of this technology will be limited.
While this type of advanced technology will most certainly be beneficial to its users for countless reasons, it is important to note that “mechanical” human-voice interpretation cannot be applied across the board. In fact, for many important tasks, this would be disastrous. The notion that litigious depositions could be recorded from a natural speech-recognition device is nearly as dangerous as relying on the Internet to diagnose the symptoms of a heart attack. One cannot hang such importance on a machine now, or in the future, because of human intuitions that drive our most important decisions.
For more, read the following articles on a range of topics from various uses to court battles regarding the new voice-recognition technology: