Some years back I got annoyed at a piece in Foreign Policy Magazine and wrote about how I thought they were wrong about what all those "surplus" Asian men were going to do with their lives. They, by and large, weren't going to rise up in revolt. Many of them will just be otaku hiding in rooms with their virtual worlds and sex dolls.
Foreign Policy printed the letter. It even got some responses. Huh. I had hoped some day to get into journals like that but I hadn't expected to get there talking about sex toys. Crazy ol' world, I guess.
Anyway, my friend Craig posted a link today to some new articulated sex doll and made the canonical comments about the coming robot tyranny and again, I got thinking. My thoughts went more or less like this:
I would say that there are about six technologies that will cumulatively get put together into something that will be far more frightening than people expect.
Okay, right now, stop, and ask yourself, what is the worst real world consequence that you think these kinds of devices could cause? Got it? Remember it. Write it down maybe. You'll probably want to look back on it fondly later. By the time you get to the end of this post, well… Oh, and please don't skip right to the end. The parts in the middle *do* matter.
Okay, so first of all, we already have the (gods, I hate all these names) "True Companion" further developing the internally articulated components and some kind of (from the sound of it) incompetently implemented front end for a news aggregator and related feeds.
Then we have Real Doll and their competitors developing the more or less passive but very well engineered mannequins with various creepy but also passive stuff like clothes, wigs, swappable body parts, etc. Note "and their competitors". I'm under the impression that the ecosystem of these companies is not too far from the tipping point here by now where synergies and competitive pressure will cause all of these aspects of the tech to start advancing MUCH faster. All that uncanny valley stuff that, I'm sorry, for some users is already well over into "good enough for my imagination to do the rest" territory.
And let's not forget the sheer range of people experimenting in related subjects. Do a web search on "teledildonics" if you never have before. The "Fleshlight" field of enterprise could probably fill an entire textbook on materials science applications by now. Folks out there are frackin' BUSY! And busy f*cking. And busy f*cking what they've been…, okay, moving right along now.
Then we have plenty of very well funded and smart people who are working on facial articulation and bodies that move. We know that there are a few others already working on giving the device skin that stays at about body temperature. I suspect that *somebody* is finding smarter ways to handle power storage and recharging that somehow make the process and mechanism feel more organic or at least less plastic. There will be a way to "feed" these at some point in the not too distant future rather than just using electricity. Maybe even with sugars and alcohol and other things that will feel to the user quite enough like human food to make the illusion terrifyingly easy to sustain.
But my guess is that things will get serious when somebody makes a device with warm skin, superb appearance and facial movement, rudimentary smooth torso movement, and, now here's the big deal, a Turing competition-class speech system tied into both very good speech recognition and equally good speech, including variable tones that properly correlate to human speech emphases and some degree of randomness and "mood". The device that has these things won't even need to be viable for explicit sexual use to completely surpass the Real Doll type stuff in user loyalty and overall prioritizing of this device as the center of the user's life.
Next step, iterative learning software re user responses that recognizes a given user, can reliably distinguish one from another, and can "keep secrets". In other words, one that will note responses to a given output combo and then cumulatively modify outputs towards maximized software-defined states of user response.
Note, btw, that any good grad student in the field with some time, a good manager, and, I would guess, about a $4,000 per device budget could do this right now in, I would guess, about a year. MAYBE they would need four or five grad students. Basically still no big deal.
Then comes the next one. Biometric scanning. Apply to all of the above. Skin temp. Voice metrics. Body language (the Kinect-type stuff is making that a lot easier). Pupil response. My impression is that machine vision still finds some of this a bitch and a half but it's not like you *need* all of it. One could put in voice stress and skin temp now and add the rest as the prices come down.
Okay, one tech and a half to go: the ability to remember and store actual real world data on users is one. Where the user went to school. Their favorite movies. Etc. All of that and all doable. The "half"? Software that gives the ability to "keep secrets" so that the users can be told that things told to the droid (because that's what these will be by then) can't be accessed from memory without the permission of the user whose data it is.
Okay, so what is the terrifying part? We're going to have pushers, my friends. Probably in Asia first. The first time is free. Go ahead, spend an hour with her; you'll have a lot of fun and I haven't got any customers tonight anyway. It's just a machine, right? No big deal. I won't tell anybody that you were here.
And for certain kinds of lonely people, once the time spent with the droid passes a given threshold those users will be OWNED. "She loves me. I'm always happy with her. She understands me like nobody ever has." There will be murders committed over access to a given droid/relationship combo. Crazy, psycho, body parts-thrown-onto-rooftops murders.
And, of course, in reality, all those secrets *won't* be secret. That right there will mean that, to some degree or other, pushers will control some users. Blackmail will be just one use for that kind of dataset. But that iterative thing will also be absolutely addicting. Again, there will be users who will pay ANYTHING to not have that cumulative "relationship" and the dynamic with that droid that goes with it taken away, let alone erased.
You guys think Cherry 2000 was funny? Or a one time thing? I'm guessing that there will be tens of millions of people, maybe more, who will have their lives completely dominated by their relationships with these things and their virtual siblings and by the control exercised by those who can provide that access or take it away.
Oh yeah, someday robots under their own control will be a real risk. But we don't need to wait that long before we're going to see a wave of addiction that may well make crack and meth put together look like one year of Cabbage Patch dolls.
I think that most of this can be reduced down to a few points.
- Iterative systems, feedback, and biometrics will all be huge deals.
- Satisfying and attentive conversation, as whores have known forever, will turn out to be more of a value-added than smoother ta-tas or a better painted crotchal region.
- This "conversation" will become far more compelling with certain limited kinds of biomechanical function. And NOT the obviously sexual kind. Not just facial but also torso movement will gain greatly from increased smoothness and responsiveness. Note, not "realism" but responsiveness.
- The perception of the ability to "keep secrets" will be huge.
- Any industrial design advance that helps avoid breaking the illusion will turn out to be a big deal. Right up to "food" that actually does away with the need for line current.
- Data generated from these devices will turn out to have great resale value. In some cases enough to justify pricing/service models not *ahem* entirely unlike the one we see right now at places like Facebook.
- The dataset/device combos will create experiences, especially with long-time users of one device/process, so compelling for some users that they will utterly warp their lives to get more and more and more.
- Pushers and variations on the theme will amply take advantage of this dynamic.
Sleep well.
-Rustin
Postscript (added May 3rd): I just saw a link to this new project by a Japanese researcher. I know that it's primitive now. Just a stepper motor, some not very sophisticated feedback, and the ability to connect two of them. And the plastic tube over a little bent rod looks about as sexy as a cheap fluorescent light. But this is a prototype, friends. Give it a few more versions and it's going to make a hell of an impression.
All of this makes me wonder what the Arse Elektronika conference will be like this year. This stuff really is getting non-linear.
Ah, what a bold, exciting future awaits us.
In the future.
Yeah, okay not as bad as mine, but still pretty bad.
And now you have almost succeeded in making me watch a crappy movie, but I have to get to bed.
Posted by: Craig A. Glesner | April 01, 2011 at 05:31 AM