Microsoft downplays Vista's speech recognition exploit

Redmond (WA) - Microsoft’s security bloggers are downplaying a recently discovered vulnerability in Vista’s speech recognition. A group of hackers have been able to delete files by playing an audio file on the victim’s computers. Microsoft says the hack is not a big deal and adds, "there is little if any need to worry about the effects of this issue."

Click through the Windows Vista launch gallery ...

Vista’s voice recognition feature is mainly meant to help the physically disabled operate their computers. Text can be entered into word processors simply by speaking words and commands can be given by saying certain phrases. For example the typical "delete" and "select all" commands can be sent by saying those words.

The vulnerability involves tricking the victim into playing an audio file that will give commands to delete or modify files. The audible commands would be picked up by the microphone and then acted upon by the operating system.

Microsoft contends that the hack is difficult to perform because the attackers have to bypass several barriers. First, the speech recognition must be turned on - the feature is off by default. In addition, both the speakers and microphones must also be turned on. Of course, any voice file being played back would also presumably be heard and Microsoft says the victim could easily stop the file and prevent any damage.

While the hackers have been counting on Vista’s speech recognition accuracy, Microsoft representatives have had much less success in demonstrating the feature to the public. In this infamous video Microsoft blamed ambient noise on the operating system’s humorous interpretation of the spoken word.