The author of the webcomic Left Over Soup proposed a sexual equivalent (or parody?) of Searle’s Chinese Room argument, posing some interesting questions about what it means to have sex, consent and relationships if there is technological mediation:
Imagine a complex assembly of vibrators, hoses, and mechanical arms, designed to stimulate a human body in every conceivable fashion. A pornstar shows up, strips down, straps on a variety of sensors (pulse, temperature, skin capacitance, genital response) and grabs the control stick with three buttons labeled “I like this”, “I don’t like this” and “No, hard stop”. The pornstar assumes the position and gives us the thumbs up.
Now, in a separate room, a gamer is shown a series of controls and monitors, all displaying entirely arbitrary abstract symbols. As the gamer plays, despite not knowing what the various pings and gradually filling bars actually signify, they begin to become proficient – this sequence of commands followed by that sequence gives a good result! – and the game continues on and gets more and more fast-paced and complex until – fireworks! – a climax is reached.
Has the pornstar had sex with a human being, or was this merely a form of masturbation? Has the gamer – who never realized their game had a sexual connotation at all – had sex? If so, although the pornstar’s consent is obvious, the gamer never consented to be part of a sex act – was this a form of sexual assault on the gamer? If we record the inputs the gamer made and play them back again and again (assuming the human subject reacts the same way each time), is anything different about it?
Information has been transmitted both ways, and reacted to. Did communication take place? Presumably, the pornstar would notice the difference if a different gamer was playing, and the gamer would notice if a different pornstar was performing. Can these two people, therefore, be said to have a relationship?
Sex with systems
“Having sex with a human being” can both be a statement of fact or a statement of intention/knowledge.
That the sex is remote, using various intermediates, does not make it non-sex: I think people would agree that full-body bondage gear sex is sex, despite the lack of skin contact. Phone sex and chat-room sex are also forms of sex, without any physical contact. Dibbell’s classic essay “A rape in cyberspace” makes a good case that text-based interaction can be a sexual medium, and that even non-physical interactions may require consent.
However, there is a lack of sexual feedback to the gamer. There is informational feedback linking the people, but only one is getting a sexual experience (assuming the gamer does not have a very unhealthy attitude towards scoring points). They are certainly not having sex with each other, and hence I think it makes sense to say only the porn star is having sex.
That neither knows about the other seems to preclude any claim they are intentionally having sex with each other, or that they have a relationship. This is not too weird, people in the same bed may do sexual things in their sleep they might truthfully disclaim any conscious intentionality of: the whole person is not involved, just various subsystems. While neither has consented to sex with the other, both have consented to the situations they are in and are fine with it… for the time being.
So I would be arguing that what is “really” happening here is that the porn star is having sex with a system. The system includes the gamer, but only as a part of the system, just like the original man in the Chinese room. The gamer is not having sex, just as the Chinese room man does not know any Chinese. This is very much the standard systems reply to the room, applied to a bit more salacious problem.
Searle’s point about syntax not being enough for semantics could be paraphrased into the claim that erotic interactions are not enough for having sex with someone. I am somewhat reminded by Rea’s analysis of pornography, where he discusses the issue of how a shoe catalog might be erotic to the right kind of fetishist, but cannot be said to be pornographic. By this account the porn star is not having sex with anybody.
Retroactive sex and mediated interaction
Being wrong about the existence of morally relevant outcomes of actions is also nothing new. If a boy plays a computer game that actually controls real space fleets fighting, he cannot meaningfully be said to be responsible for their fate even though he is causally in charge. To be morally responsible one needs to understand what is actually going on, not just what seems to be going on. Note that this can lead to trouble if such understanding arrives: the gamer and/or the porn star might be horrified if they learn the true setup, just like the boy would be shocked to realize his actions had enormous consequences. While they might try to hold themselves blameless due to ignorance, they will know they performed actions that mattered and now hold extra emotional and social meanings. In a sense the couple will now have had sex with each other.
Knowing about the true nature of the system you are interacting with may change the moral meaning of the interaction, sometimes even retroactively.
We are more and more interacting with the world through complex mediation systems. When I post this essay, it will be sent through a vast infrastructure to readers; the layout and even language can be changed, and potentially intervening software could add adverts or censor it. Reader responses will similarly be filtered through the system, including spam filters deleting certain posts. Many parts of the system are unknown to either of us. In this case the existence of an author and readers remains transparent, so the main ethical issues of the mediation might be if it does something unacceptable with the communication: we understand enough of the mediation to take moral responsibility for what we use it for.
However, a system can be opaque: the moral agents involved are no longer visible to each other. It is a matter of degree: Wikipedia editors can be found, but most of the time it looks like it magically updates itself. An autonomous car might have code acting as a moral proxy for a designer.
A teledildonics entrepreneur I once met discussed what to do with the data streams from his customers; having a “replay” option seemed doable, and I suggested some data mining methods to try to improve the signals. Were the users aware of this, I assume there would be no issue. But it is easy to envision one partner having a headache and sending a replay, or that the software introduced enough change that one might say the system was now a sexual participant. One way of analyzing these cases is to consider the retroactive response of finding out the truth. In the headache case there might be some real upset at the deliberate breach of trust, while the improvement case might depend more on personal views. Indeed, a recent court case involves a customer upset that her app-based communicating vibrator transmitted rather sensitive information to company servers; had she known this, the lawsuit states, she would not have bought it.
Demanding transparency might not be the right thing: we do not want to know every detail of our complex lives. We want to know the morally relevant parts: who are the other actors involved? Are there automation acting as moral proxies for someone (e.g. censorship software)? What are the morally relevant outcomes of our actions? Ethics aware design should aim at building systems that help users know the answers of these questions in a relevant way.
We don’t want to wake up with the wrong kind of entity – whether a system, a human, a company, or spambot – in our bed.