The Year of Speech Recognition

Is this the year of speech recognition? GigaOM thinks speech recognition is getting ready to take off — Is Speech-Recognition Technology Finally Ready for Prime Time?

IN 2006 AnandTech had similar headlines — Speech Recognition — Ready for Prime-Time?

CNN called it in 2001 — Speech Recognition Technology will hear you now — “After years of hype and false starts, automated speech recognition (ASR) technology is ready for prime time”

SpeechTechMag made the claim in 1999 — Speech recognition technology is “ready for prime time”

Apparently Comdex buzz in 1998 was that 1999 would be the year of speech recognition. Again.

At one point in the mid/late 90’s, the Microsoft Speech Recognition team worked for me thanks to a reorg. All great people, very enthusiastic speech supporters, and convinced that widespread speech recognition was right around the corner. I loved the team but it seemed apparent that the hardware and software were both generations away from achieving useful speech recognition and so ultimately I worked to get the team moved to Microsoft Research.

So call me a skeptic. Over the past 15 years, you would have done well to always bet on “No” as the answer to “Is this the year of speech recognition?” I don’t see a lot of reason to change the bet. Speech is tough, we are demanding, we expect a system to respond to conversational input — rapid, lots of colloquialisms, speaker-independent, etc. Yes there are systems that can respond reasonably well to very limited inputs but as a general use technology, I’m not yet optimistic.