Before getting started, I’ll let you watch the video, and then discuss it afterwards:
My first reaction to this was to wonder how much of the scene was scripted. Even the most developed interactive bots still struggle to put a comprehensive sentence together, let alone with the addition of emotion and gesture. Though MicroSoft must be taking this seriously or else the whole project is a waste of time. So my conclusion is that yes, we are looking at real A.I. when we look at Milo, but chances are, the A.I. created so far has been specialised - and is probably only adequate enough - for the purposes of the demo we’ve just watched. Ask him to comment on the weather or your make-up, and well, dare I say a BSoD probably couldn’t be more wished for as a convenient way of ending the awkward silence.
But let’s look at the bigger picture. When I say this, I think, how BIG is this bigger picture? This could end up in numerous places. Imagine a PC game based in a medieval of fantasy setting, similar to Oblivion. You’re about to enter a forest on a mission to rescue some damsel in distress, but you know the forest is teeming with dangerous wildlife, and bandits. At your side, you have not just one squire, but a small band of brothers. Being able to give orders verbally without having to click generic, limited options is one thing, but having to use emotion and gesture to give a pre-battle motivational speech which could entirely sway the mood of the battle, or even convince nearby eavesdroppers to join in (or if done badly, scare them off), that would bring gaming to an whole new level never even imagined.
But where next, outside of the gaming world? A boy that could actually genuinely tell you that your dress looks good and recommends what colour shoes to go with it? A virtual teacher to take over the job of substitute teacher in nursery? A.I. of this level may also have its therapeutic qualities, I’m sure there’s an opportunity for helping autistic children somewhere here.
But I’d say if there was a line, this would be it. The social consequences of anything further could be disastrous. An A.I. bot capable of doing all of the above, would be very human indeed, I can imagine youngsters in a few generations time being brought up unable to distinguish humans in the real world from characters on their TV screen. Loners would become more lonely, as they no longer need friends at all, when they can just load one up from their hard drive.
If your onscreen friend insulted you, you could just delete him, or if you’re tech enough, could reconfigure or script him (or perhaps if the AI’s sophisticated enough, just tell him, and he’d never even accidentally do it again). If you’d like a new friend with a specific personality, you could make one on the spot, and then when you want him to go away, he would do so without asking “why, what’s up?”.
Now imagine the frustration of not being able to do this with real-world friends. I’m starting to think perhaps this could backfire and turn non-autistic people autistic. It seems that with mankind’s ongoing quest for perfection, we’ve given up altogether and moved on to create a super-species which could, in generations to come, ruin everything it is to be human.
And who knows, but why not, while on the topic of humanity… a character capable of performing virtual sexual favours. I wouldn’t be surprised if that was the aim of the project after-all; seriously, who could possibly work at MS, on a project such as this in particular AND have a girlfriend? Think about it. If such A.I. ever went commercial, there is no doubt in my mind that there will be a whole load of characters to chose from, as-well as functionality to design your own.
Well, I probably shouldn’t say such things about the team behind this project, and probably speak out of jealousy, as I honestly wish I was there working on this project with them. When I complete my masters in artificial inteligence, I think I found the place where I want to go. This is the current apex of AI technology, without going as far as to rely on robotics.