A recent Windows Vista voice recognition demo goes horribly wrong:
Several tries at making the computer understand the simple salutation “Dear Mom” was read by Microsoft software as “Dear Aunt, let’s set so double the killer delete select all.” Attempts to correct or undo or delete the error only deepened the mess.
You can watch a clip of the demo (why does the reporter in that CNBC clip feel the need to explain his bad joke?).
I thought this was sufficiently humorous in an "all your base are belong to us" kind of way that I whipped up some t-shirts at Cafe Press.
That anyone thinks we're going to be talking to our computers all day long is a little baffling to me. When it becomes possible to talk with our computers, we may, but talking at them is just plain awkward.
How often are you in an environment where it feels natural to speak to your computer? Would you dictate an email or a report while sitting in your cube? What about controlling your computer? Do you want everyone to listen in when you say "tab tab open tab tab tab open tab open project dot pee ess dee"? Even if the software was smart enough to recognize "open all of the files related to my current project," do you really want to be in an environment where people are issuing verbal commands to their machines all day long? No, of course not.
The exception is if you are not free to use your hands at the moment. Such conditions occur while operating a car or some other machinery, or perhaps if you are disabled. Or maybe it is inconvenient to use your hands because the activity is directly coupled with speaking, like when using a cell phone. In these cases, I can see a compelling reason for designing sophisticated voice recognition interfaces.
But until the technology arrives at a point where we don't have to repeat ourselves five times, or speak using the Voice, without the infernal machine turning our words into bad poetry, it's all a bit silly.