At Law Technology Today: Talking to Yourself: Is Speech Recognition Ready for Prime Time?
Apparently not, as this article is only available as an audio podcast, not as written text.
I think speech recognition is more than ready for prime time. In fact, more than half of the content you see here was dictated by the use of Dragon Naturally Speaking, the program they reference.
By my experience, the program is about 90 to 95% accurate, which is pretty good, except that the mistakes are astonishingly frustrating when they occur, creating the impression that dictation is no faster than merely typing it. But let there be no mistake: the proper use of speech recognition is unambiguously faster than any method of typing yet devised.
Even with those mistakes, I can go on for pages and pages and pages without end, writing and revising and — best of all — seeing exactly what I am working with, a luxury unavailable to those who dictate for later transcription by a secretary or service. Yet, I rarely become tired of writing. In contrast, when I type, I can feel myself slowly becoming depressed as I think of each new paragraph that I will have to type, as my hands and fingers start to recoil and my breathing becomes labored. If you’ve typed a lot, you know what I mean — it’s exhausting!
Frankly, the question is rarely whether a particular technology is ready for prime time. The questions are: can you see the strengths of the new technology, and can you adapt to leverage those strengths?