Regardless, just using Kinect as a microphone and speakers won't cut it for me. Some people have a lot of family and a Kinect would pick up a lot of stuff that people don't want broadcasted. That's one reason why headsets are nice, as the mics wont capture that much other sounds except the one from your mouth, while the Kinect mic is obviously able to catch normal speech from several meters away. Also, in fast paced shooters for example, many people prefer to have headphones to better hear directions and small audio details (such as footsteps).
That said, I was impressed by that Kinect vid and I'm sure that it does the job well baseline, but it's not really ideal for online gaming IMO, unless you live alone.