Not necessarily. You don't need full speech synethsis to do what they have shown - it just needs to pick up certain key words, and be able to say certain key phrases. The entire demo coulda been done simply with a big bank of ready recorded phrases

Also, I have proof santa doesn't exist - I stayed up all night and saw my dad stealing the beer and mince pie we left out for santa, yet he still got presents - So Santa either doesnt exist, or at least isn't omniscient

Of course it could be done with a big bank of recorded phrases. That is exactly how one would go about creating a scripted demo (whether there are a few possible alternatives or an entirely on-rails experience).

Speech recognition in itself sucks in its current state, on top of it add gesture and face recognition. It's going to take a decade before we see anything like that demo in a real game.

The demo was scripted, thats not to say they cannot expand upon a concept they showed last year though.

I dunno. I don't see the amazingness of this. And to say SF writers haven't thought of it is PR BS. OK strictly Molyneux is right because what SF writer has ever thought of using an AI character that only resides in an immobile TV screen? No all SF AI characters are cyborgs or the like. So I guess Molyneux gets out off on a technicality there.

And I don't see this as being a major showcase for Natal in and of itself. Any camera and microphone set up that has the appropriate resolution/fidelity can be used for voice, face and image recognition. It's the software grunt behind the sound and image inputs that does all the real work. Are people suggesting there's nothing else out there with the processing capacity to handle the necessary software? Dare I say it but Eyepet does the equivalent of the fish drawing then scanning thing (Milo appears to improve on it, but if anything looks fake about the video demo it's this) and you can teach your Eyepet to "sing" (which is sound and pitch recognition, and there is already well established word recognition and response software out there, which again Milo builds on but certainly doesn't establish) and that is a rather cheap game with not much image and sound recognition grunt behind it.

Milo is really putting established and available interactivity together in one package. That is cool and impressive for sure. But amazing and ground-breaking? I don't see it myself. Milo is evolutionary, not revolutionary. Hence I don't really think it is so far advanced that the E3 demo had to be entirely fake for the reasons people think. The video demo was probably tarted up because Milo was not quite able to do everything in the demo in real time as smoothly as was shown in the video. And obviously the interaction was scripted so one can never know whether Milo's responses were equally scripted or extemporaneous, e.g. did Milo have the "choice" to either say he'd done his homework or not done his homework, or was his homework completion status a set parameter?

@ scottie: Santa isn't omniscient, he's omnipresent. Oooh dub-le entendre.

No they don't, he just needs to recognise key words. Seaman did that on the Dreamcast a decade ago. This is just the next generation of that. Peter has said before, Milo is designed to trick you into into believing that he is more then what he actually is; a virtual pet with a few image scanning gimmicks.

This demo is also impressive. Look real people with AI interaction on video.. must be real. Is this the MOVE next gen!? OMG! Nothing short of amazing! Good game, well played!

http://www.eurogamer.net/articles/e3-project-natals-molyneux-and-milo-interview?page=2

http://www.develop-online.net/features/557/About-a-Boy

take a look at these interviews.  very good reads about milo.

“At Lionhead, we’ve always been fascinated by AI and using AI in gameplay. A lot of AI is academic research; funnelled down neural nets and learning – it’s very antiseptic, there’s no emotion behind it. It goes back further than Lionhead, back to Bullfrog, where we were playing around with these simulations of little people and their little minds. Emotional AI is not real AI; you couldn’t write a paper on it. It’s more about how you use weak learning to make people think there’s something going on there.”

 

Eurogamer: You said he only understands certain words. So presumably you can't have a conversation about the situation in Palestine?

Peter Molyneux: The number of words he understands is built up over time. For Claire [the lady who demoed a conversation with Milo during Microsoft's conference], it's something like 500 words.

But we haven't cracked the real problem, which is him understanding the meaning of it all. He'll give you the illusion he does that. The interesting thing is you can only talk to him when the Talk icon appears at the bottom of the screen. That's when he's listening to you; the rest of the time, he's not. He's listening to you because there's a context in which you can talk to him.

One of the journalists who came in before you had obviously read up on the Turing test. He asked Milo one of the questions in the test - 'Do you remember when we met yesterday?' Well, of course, we haven't cracked the Turing test. If we had, then applying it to a computer game would be the last of the solutions we'd use it for.

god help me if he really thinks the public is going to spend time talking to in imaginary video game character.

and i feel sorry for the lonely people who buy this and consider him a friend.

I'm sure some moron will make him a facebook page too.

I love how some people here are talking about speech recognition and speech synthesis and facial recognition software like they are experts in regards to the various software. I have a friend who is blind and uses speech recognition to send emails. He says it works wonders for him, and that it was very good at picking up the exact words he was using, and rarely made mistakes. I haven't even talked to him for 4 years, so maybe the software is even better. And he's Japanese and has a horrible accent, and he still said that the software he used was so good that it barely made mistakes.

Granted, the software might be more expensive than more standard speech recognition software, but the technology is surely there. Also, there's more to speech recognition than just the software, but don't forget the microphone quality is important too. I don't know about you, but even I have trouble understanding some of my friends on the phone from their shitty micss, so how could you expect a computer to pick it up properly sometimes? Hell, I have trouble understanding some of my friends when they just speak sometimes, not everyone speaks equally succinctly.

Then in terms of emotion/gesture recognition, at least in terms of the face, all you have to do is set a baseline. Which is why Molyneux said before in some interview when someone tried Mile to smile and then frown, since you then set a baseline for both. It shouldn't be too hard to store those images and try to match them up at times. I don't think the software will see and understand each and every little thing, but it's a start.

But I'm excited for Milo. The implications of the work done on Milo are staggering. Imagine playing an RPG where you can actually play as a person and interact with people. It'll take a while until the A.I. truly works the way people want it to, but again, it's a start.