The art installation was set in a dark room. Patrons could freely walk around the exhibits, upon podiums - which were spotlit from above. The two podiums contained a six foot by six foot clear acrylic box. Inside of those boxes were human beings - one male, one female - devoid of clothing - straddling a white - medical looking - fake leather padded saddle. Their heads emerged from the top of the box, and their arms rested on arm rests to either side of them. In front of their faces, a screen - with constantly scrolling information - text, images.
Their skins were dotted with various contacts - a line of round self-adhesive contacts along their spines, especially around the back of the neck. They both wore headsets with sensors positioned around their skulls. They both had contacts attached to their torso, several around and on each nipple, and their public regions. The male was outfitted with a fabric holster into which his penis was placed, which served to hold several contacts in different places. The female had a transparent plastic plate with several contacts on it held against her vagina, held in place by elastic cords around her waits and each thigh. Although it wasn't clear from the outside, they appeared to both have something in their anuses too, as more wires emanated from there.
The hundreds of wires from all of these contacts were bundled together and went to an interface box. The interface box was connected to a computer, which sat at the edge of the podium. A screen showed a graph of hundreds of signals which varied over time. The humans both stayed quite still for much of the time, but sometimes flinched or twitched. When that happened, the graphs showed various spikes across the signals they were displaying. The largest graph on the display was labeled "AROUSAL LEVEL" - the line for this was mostly flat and at a level of 98%.
Integrated into the saddles, directly under their genitals, was a small funnel. The funnel was connected to a vacuum pump, which periodically activated to remove any buildup of liquid. The liquid was pumped into a clear, thin tank with measurement markings on it by the bottom of the box.
Patrons walked around these exhibits, observing their predicament. The humans did not acknowledge the patrons - partly as they couldn't see them in the dark room, but mainly because they were lost in their own world of frustration.
---
The artist's name was B-Twenty. They had changed their name in 1998 after an personal epiphany in which they decided that they aimed to transcend the human experience, and the best way to start that process was to give themselves a name that didn't fit into any normal society. They also publicly declared themselves as gender-neutral at this time - although they had always been this way - it felt like the right time to cement their new identity within the community. Over the years, they had loosened up on the name and were happy for their friends to call them just B.
B had recently become intrigued by the modern trends in computer driven software development - two aspects of it, in fact. One, being the intense desire to drive attention to the detriment of everything else. And the other being optimization of a businesses goals using AI. After talking to many academics and professionals in the field - and indeed becoming part of the industry - they had begun to understand that the danger of AI was not the apocalyptic "grey-goo" scenarios that fascinated science fiction - in which a maligned superhuman general intelligence destroys all matter in the universe to optimize some outcome - but rather the more imminent threat of much lesser intelligence enslaving the human race by exploiting their own desires.
That's how they came up with the Arousal Farm. In a parallel to the goal of many social networks or apps, the goal of the system was not to maximize attention, but to maximize sexual arousal. And like attention, sexual arousal is indeed a desirable thing - in balance to all other activities in a person's life. But if the goal is just to maximize a simple parameter - there is no balance between that and other parameters (like social interaction, companionship, new experiences) that a person needs in order to live a healthy or satisfying life.
B worked with a number of people to build this project. Several engineers - medical, electrical, and software. They had to consult various people on ethics. And not least, the models themselves who had volunteered to be the subjects.
The system was constructed in two parts: the human computer interface, the AI. The interface was two way - it was possible for the computer to sense various parts of the human's body, as well as stimulate them directly or indirectly. Some of the sensors needed to be labelled so that AI knew they had a strong correlation with the arousal of the human, but most parameters were unlabelled and it would need to determine the relationship between them on it's own.
The process or learning was fairly complex - just like the algorithms to maximize eyeball time on adverts. At first, it would try to build a model of the subject. It would watch the sensors when there were no stimuli - then systematically try different combinations of stimuli and measure the response and encode that into a neural network. It would then repeat this process several times, each time improving the predictive power of the model.
After creating a model of the human, it would generate a number of new algorithms which would be tested against the simulation. Each one would be scored by the overall level of arousal it achieved over a fixed period of time. The one which was the most effective would then be used as the basis of the next round, with a new set of algorithms each of which are a slightly modified version of the last. This was an implementation of a genetic algorithm.
However, the model of the human will not be perfect - and so this genetic algorithm will be optimizing the arousal of the model, not the actual subject. So every few hundred generations of algorithms, the best algorithm is tested on the real human. The results of this are used to further refine the model, which is then used to test the next generations of algorithm.
This process repeats over and over - but the computer is fast enough to be able to perform trials on the human subjects every minute. The monitor shows the progress of the genetic algorithm, and which algorithm is currently being trialed.
The approach to building this was challenging for B and their (open minded) engineering colleagues. Humans are notoriously difficult to predict, and especially in this regard as their responses to stimulus like this are non linear. One of the key things that the algorithm would need to learn to deal with is the orgasm. If it accidentally causes an orgasm - whilst the arousal level will spike briefly, it drops steeply afterwards - an undesirable outcome in optimizing the arousal levels over time. For this reason a manual orgasm detector was built, which would reduce the score of the algorithm by a significant amount of it occurs. But in the early generations, it will have never seen the human have an orgasm, so the model will not be able to predict it accurately. Only after having seen the orgasm occur, probably several times, will it be able to accurately simulate and predict when it will occur.
By the time this system was built, B's team was confident that it would perform mostly as intended - or if it did not, it would at least be interesting to see how it failed.
---
Tiffany had been a model in the art world for a few years now. At 22, she was both glad for the extra income and for being part of the art world she loved so much. In both ways, it allowed her to continue and further her studies into becoming an artist in her own right some day. She had always been confident in her own body, and didn't feel embarrassed by others seeing it without clothes. Her family were naturists and as a young girl had often gone to naturist resorts on holiday with her family - and that had given her both a respect for her own body and other peoples. She appreciated how different every person is, and judging someone for the shape or size of any part of their body didn't even enter her thought process - and the people around her during those breaks were no different. She also had no connection between nudity and sex. She imagined that many people had that link, and therefore associated their naked body with their sexual performance - which often induced anxiety. She felt lucky not to be burdened with this. That confidence made working as a model much easier to her than perhaps some of her peers, especially on jobs which required nudity.
