Back at Microsoft’s E3 press conference, Peter Molyneux introduced us to a charming little artificial boy named Milo in order to demonstrate the capabilities of Project Natal. We came away very impressed at the AI’s ability to recognize human speech and movement.
However, now that respected game journalists have been able to interact with Milo themselves, a little bit of doubt is starting to sink in. According to Gametrailers editor-in-chief Shane Satterfield, those chosen to interact with Milo were told to do so in a manner precisely in step with the demo shown to the public. Even more disturbing was that Satterfield believes someone was controlling Milo behind the scenes. He goes on to say that the Milo demo is “not legit” and “there were a lot of smoke and mirrors going on there.” Yikes. Not exactly the glowing review we were hoping for.
It seems that the deal with Milo is that he reacts to the tone of the player’s voice rather than actually understanding what he or she says. Also, Milo has been programmed to speak in vague terms so that he makes sense no matter what the player says.
The good news is that this is just a demo. If it was ready to be played with by the general public, it would have been out there on the show floor at E3. Maybe some day in the future we’ll have some real voice recognition in games, but it doesn’t seem like that time is now. Still, it’s nice to see the effort to innovate taking place.
Better news is that Satterfield didn’t have any problems with the Ricochet demo, which looks like a lot of fun.