Ruminations on AI and ASFR

[D.Olivaw]

Hi everyone,

So, I was watching G4TV the other day, and came across mention of the author (I think the name of his book was "Love and Sex with Robots") who predicted people would be having sex with robots within the next five years and meaningful relationships in the next ten. Here are my ruminations on that subject.

If there are two things that can drive a field or application forwards, they are that it either has a potential military application or that it has a possible sexual usage. I think AI technology fulfills both of these requirements.

Most of the world's modern militaries are betting on unmanned vehicles to make up an important part of any near-future fighting force, from UCAV's (Unmanned Combat Aerial Vehicles) to unmanned surface ships and submarines to direct battlefield applications in the form of robot mules like "Big Dog" or fighting platforms like "Crusher." At the moment, all of these systems are purely remote controlled but some (like Global Hawk) are autonomous in a big way. It is natural that as these technologies develop we will seek to increase the native intelligence of the individual platforms so that they will need less direct human oversight, allowing them to operate as a potent force multiplier. Perhaps in the end this will lead to something in the pattern of the "think-tanks" from Ghost in the Shell. Human oversight would basically consist of one human "operator" overseeing several autonomous platforms, acting only to relay high-level tactical commands and authorize the use of deadly force. It is a relatively safe bet that military applications are going to drive AI technology forwards.

Humans have been devising sex-aids for as long as there have been humans; some of the earliest records we have of such devices are from Greece, where dildo-usage was apparently so common among upper-crust Athenian women that there was a popular comedy written about it. The isle of Lesbos (three guesses as to what modern word that name gave rise to) was generally recognized as the love (and lust) capital of the Greek world, being the source of both the highest quality toys and the highest quality courtesans, both male and female. Indeed it was here that Sappho penned her famous erotic poems. In the age of sail, French and Spanish seamen apparently constructed "Dames de Voyage," primitive sex-dolls made of sailcloth and used to eliminate their pent-up urges. It is also rumored that slightly more realistic contrivances were used aboard German and Japanese subs during World War II.

I don't think it is a stretch to assume that we will couple growing skill in robotics and AI with our sexual desires to produce robotic sex-dolls in the near future. As they become more realistic the industry will grow, as will the demand for ever more lifelike dolls. This will certainly push forwards both robotics and Artificial Intelligence as owners begin to seek companions rather than just masturbatory aids.

These two forces, martial and sexual, may act on the one hand to increase the ability of AI's to interact with a complex, ever changing environment (the modern battlefield) and to be flexible and decisive in their planning while on the other hand greatly improving their ability to interact with human beings. As these two vectors in AI development merge, what will they produce? It is impossible to predict what roadblocks we will come across as we advance, or what new innovations and technologies might speed us forwards. Currently, human programmers have trouble fully concieving the way their code will work, so complex has it become. How will we program machines whose computational power will dwarf that of today's supercomputers? It may be that humans will only write parts of the code, that much of it will be constructed using genetic algorithms; or that a vastly more powerful language will be developed (along the lines of the Visual Programming Language in Ghost in the Shell) that will allow programmers to more easily grasp the nature and form of their code. It might also help if we learn how to augment our intellectual capabilites with cybernetics.

Will the AI's produced by this process be traditional weak AI or will they actually become strong AI? If a machine can interact with us and the world on the same level as a human being, is the distinction even meaningful anymore? How will we determine whether the intelligences inhabiting the intermediate steps of this process are deserving of rights? If so, what rights?

For my final questions, I'd like to ask what technology you think would be most likely to produce a truly sentient being, rather than just a weak AI: traditional programming (Turing Machines), current digital neural nets (also Turing Machines), or analog neural nets (each artificial neuron can be at any of a number of states, not just on or off)? How long do you think it might be before our species creates a conscious AI?

My own answers to those two questions are:
1) No idea, though I favor analog neural nets because we know they are capable of supporting consciousness (the lump of jelly in our heads is essentially an analog neural net).

2) Unfortunately, I think the challenges are such that even with a serious investment of energy most if not all of my generation will be but dust in the wind before the first conscious AI walks the Earth (assuming human lifespan doesn't dramatically increase in the near future). I hope I'm proven wrong.

[Borias]

.

[DollSpace]

2) Unfortunately, I think the challenges are such that even with a serious investment of energy most if not all of my generation will be but dust in the wind before the first conscious AI walks the Earth (assuming human lifespan doesn't dramatically increase in the near future). I hope I'm proven wrong.

I hope I can prove you wrong on that one some day!

Catie

[rabiator]

I think we'll be stuck with weak AI for the foreseeable future:

Strong AI requires a major breakthrough in understanding intelligence that may or may not happen. Besides, emulating a complete human brain would take significantly more computing power than current computers have.

At the same time, weak AI has already shown that it can beat humans in a narrowly defined field. Consider chess computers, which can beat grand masters these days.
I think military development will continue along this path, with specialized AI for specialized robots. Like
-one for unmanned fighter planes
-one for anti-aircraft guns
and so on.

In the sex industry, big technology programs seem to be rare. Something like Real Dolls seems the maximum that this industry pulls off, so I don't expect spectacular innovations there. Of course, they will happily exploit developments by others:
If there is ever an academia project that creates realistic androids, the sex industry will make a fuckbot version

[D.Olivaw]

you may be right regarding the sex industry, rabiator; I sort of made those conclusions without too much in-depth knowledge of it. As far as your statement about specialized AI for military purposes, I think I should clarify myself. I did not mean to say that military AI would become very generalized, like human intelligence, indeed I think that what will be developed will be specialized intelligences for different platforms (like a fighter or a patrol boat or a tank). However, think of the incredible complexity even a fighter aircraft (the simplest of the proposed systems) would have to deal with. The technology would then ideally progress to intelligences capable both of more abstract tactical analysis as well as more common-sense type operations (finding one's way through a forest without being seen, etc) that can be applied to ground combat vehicles. My statement that this could lead to human-level ability to interact with the environment (for a given specialized platform) was a statement about perhaps the next century or two, not a near-term prediction.

[WinterRose]

For my own part, I don't believe these robots will be solid state affairs. Right now, Intel and the DoD are working to make Claytronic Atoms a reality. (CATOMS for short) These are essentially nanoballs of a certain size meant to be able to simulate people, objects or whatever. Without going into a lengthy explanation of the science and how they work, just imagine holodeck technology, but it's a mutable nano-solid instead of hard-light. They will change the face of our entertainment industry. As well, it may change the face of robotics. (As of 2 years ago, they were predicting having them working by something like 2015.)

Why work over and over, with fake skin, air-bladder & electric musculature, heavy endoskeletons and all that jazz when you can simulate and animate the same thing with catoms? You'll already have simulated sports events on your dinner table. Perhaps rooms of the stuff to sim whatever furniture patterns you download. You'll see it animated as people in miniature, or perhaps in full size. So if you can sim people, you can see where this is going, can't you?

Of course they WILL have to do something about heat, safety and realistic texture. I also wonder how they'll simulate a colored texture. But consider the possibilities. Will people go out? Or will they stay at home and SEEM to go out in their virtual rooms. Their actual presence in the world being a simmed catomic avatar? When people start relating Catomic avatar to Catomic Avatar, how long before they start making idealistic adjustments to their appearance? What happens to society when your very appearance becomes as customizable in real life as it is in Second Life?

It goes further than that. When entertainment shifts to solid 3d representations of the actors in question, you're gonna have to be a special kind of person to even CONSIDER going in front of a camera. It does NOT take a genius to see that when you're playing back the latest blockbuster hit with Emma Watson in 2025, someone's gonna pause the action and take a bit of a grope. The contract between actor and viewer changes when they can have intimate knowledge of your form at the touch of a button.

And of course, here's where we get into the whole idea of the prurient uses of catoms in entertainment. (How long before someone changes 'entertainment' to 'Intertainment' to reflect its more interactive nature?) If it walks like a female, and looks like a female, people will want to touch it and make love to it like a female. If true AI, or even simple VI (to use the mass effect terms for Artificial Intelligences and Virtual Intelligences) could be the animus behind an interactive catom sim meant for sex. And this is assuming people wouldn't be having sex with one another by avatar proxy using these things.

Consider it? Why have one appearance? Why limit your bot to one form when you can download the latest pirated topographies of 23 year old Milla Jovovich, 25 year old Katherine Hepburn, and Marylin Monroe from 'Some Like It Hot'? You see where I'm going with this? I really believe that this is more the direction that robotics will go in. It neatly sidesteps all the engineering pitfalls of the solid state, as easily as coded software sidesteps primitive computers like the Babbage Engine. Infinitely customizable.

Honestly, I mentioned this to the folks at World of Wonder when they were filming that episode of Wired For Sex. We actually did an interview for about an hour or so. They didn't use nearly all the stuff we talked about.

[Keizo]

Back (not that anyone noticed I was away ) This is a really interesting tech WinterRose. I remember there was some allusion to it in that old show called Wild Palms. I'm still curious as to how something like this could work though. Such as will there need to be an internal "skeleton" so that it could be structurally sustained? How do the particles receive their instructions and how are those instructions maintained? Will it be in layers to help with the illusion of things such as clothing and skin? Can an internal skeletal structure also act as a power source? Will the particles be powered by microwave and can those same microwaves be used to emit the instructions so that they can form the chains or lattice necessary to sustain the structure and create movement? Will that mess with our physiology if we come in contact?

This is definitely an evolving technology and one that I would actually prefer if it can be perfected if just for the sake of variety! Thanks for the heads up; it will give me something to think about and to research.

[stelarfox]

well, perhaps I am a bad person, but, I think an AI is still a robot, I really do not care if it can learrn, or if it's more inteligent, It was programmed, even if it was a learning process, in that case it was PROGRAMMED to learn.

so it should have no rigths, it belongs to somebody. (constructor, programmer, user).

and if at any point this machines are to be build, in my opinion they should have a number of safe mechanisms, like if they beliebe they do not have to follow orders they automatically shut down.
or a lot of overrides codes.
(so may be they malfunction and think they are free but, at the send of the override codes, they shut down, or they follow orders again , or something like that).

of course it depends on the programming and the machines can allways be hacked, so they may be reprogramed to not answer to the overrides, or something like that but if I made that kind of things it would be so secren not even the machines have knowledge of that. so, at any time they can be stoped easily.
and for the second part all the "main programming" should be in rom, (not eprom, not any other kind of memory), rom, so it cannot reprogram itself at any time.