This is my first shot at the abstract for my paper. I may be submitting to a upcomming school conference and I needed to get this together.

24×7 Chat Walkers : How artificial intelligence prostitutes are commoditizing fantasy in online chat rooms
In this paper I examine how bots, programs that are written to function as humans, are being used by the adult industry to persuade chat room participants to join pornographic websites. I demonstrate how a semiotic analysis of these interactions between humans and machines as humans reveals how identity and performative gender stereotypes influence and evolve online interactions. In particular I will explore how these bot’s user profiles and response scripts are optimized to take advantage of naïve chatters preexisting sexual fantasies, in order to fool the chatter into thinking that they are talking with another human. In order to do this I will demonstrate how the adult industry both monetizes the creation these fantasies and these human/bot online interactions. Finally, I will consider the socioethical implications of machines that “pretend” to be human. Though ethnographic interviews conducted with bot creators and chatters who interact with bots I will explore how individuals treat bots and chatters who mistake them for humans.

IRB = Institutional Review Board

Here’s the short summary: The IRB is an extension of the Government’s Office for Human Research Protections (OHRP). It’s job is to ensure that scientists don’t abuse folks. The impetus for it came initially from Nazi Death Camp experiments. However, the direct cause were a number of reprehensible experiments carried out here in the US by a number of bad eggs in the Medical and Psychological fields. The end result is that we (Social Researchers) who get any form of government funding (or are carrying out research at federally funded Universities) need to go through a lot of red tape in order to show that we will in no way harm our subjects.

Is it important? Yes. Is it out of control? Yes, at least for anthro work.

The problem for me is that online research is a very open subject area. The rules are still being laid down and that is problematic for research like mine. If all my i’s are not dotted just right I could find out that I have to scrap all my research. Conversely, if every time I enter a chat room I need to get everyone to “sign” consent forms I can’t conduct my research. Avoiding this double-bind is critical to successful completion of the project.

Yesterday, the U of Chicago IRB ruled on a number of Internet related research issues. While these don’t necessarily apply to other academic institutions, the can be used to help argue cases:

1. Any content viewable without a password is considered public
   If you can access it without signing in, it’s considered public. The general rule: if a website is listed on Google it is considered to be publicly published content. As such you do not need permission to reproduce it in studies. This includes unrestricted Blog and LiveJournal content.
2. If there are no membership restrictions for protected content it can be considered public
   As long as anyone who applies can get access to the information it will be considered public. This covers cases where a valid e-mail address is required provided that you can use a free e-mail account like Hotmail, YahooMail, or Google. As such most discussion board conversation can be considered public and you do not have to secure releases for publishing and collecting it.

   There are two possible exceptions to the above two rules:

   1. If the publisher of the content is under the of 18 there is still a need to either get parental consent or justify the waiving of it.
   2. Should the specific rules of the site/board mark content as unable to be reproduced without specific consent of the authors. The jury is currently out on this one.
3. Publicly available chatrooms are most likely public
   Sites like YahooChat, which require a free membership, most likely are public. However the U of Chicago IRB board will review these on a case by case basis. Researchers will have to address the possibility of recording information on subjects who are under 18 and will be expected to address how they will work to prevent this from happening.
4. Usernames must be changed
   Because a username can be connected back to an e-mail address, they are considered a personal identifier. Unless the individual gives specific permission to use their username it must be coded and protected as if it was a real name.

   side note: I had been concerned about how to code, considering that any pseudonym chosen could be another individual’s existing username. However the board ruled that this is not an issue providing that your notes/publications are very clear that you are using pseudonyms and that any relation to those living or dead is purely coincidental (to borrow movie jargon).

5. IRB rules do not apply to Bots
   As Bots are not human, they don’t need to be treated as human subjects. This means that no permission is required to reproduce conversations or to publish exact usernames.

   side not: Because some bots deny that they are bots, the researched needs to provide the IRB with a repeatable method for proving that they subject is an artificial intelligence (i.e. a reverse Turing test).

I’m having a difficult time finding a reader for my paper. I had expected this to be the case. Dr. Silverstein, while he said he’d be a consultant, suggested that he wouldn’t be the right person for the paper. I just talked with another person from the Anthro department and had a similar experience.

One downside to Chicago being a research institution is that even in the social sciences the faculty is very much in flux. A couple people who would have been good candidates for readers are in the field this year. So they’re out. This is another reminder that, even if I decide to do an eventual PhD, Chicago isn’t a strong contender. While this is the #1/#2 anthro department in the world, they’re strength is not in the type of work that I’m interested in. But it’s been worth the price of admission to know that I can hang here AND that it isn’t the right place for me to hang long term.

Another issue surfaced today. My project is going involved the adult entertainment industry. While I’m not specifically studying porn, they are the ones creating the bots that I’ll be analyzing. And I’ve been told that association can equal academic suicide for young male scholars. This is arguable the first time that the specter of gender discrimination has ever entered my life. The Ivory Tower apparently doesn’t look kindly on straight men and porn, even if the connection isn’t about the direct study of porn (as in my case). I’m not planning on letting this stop or worry me at this point. But I was taken a little aback by the entire conversation.

Far, far, far more important is finding a reader. If I don’t do that, I don’t graduate.

I’m primarily going to be studying these ‘adult’ bots (I’m toying with calling them sex.bots as there currently isn’t an accepted ‘industry’ name¹ for them). There are a couple reasons for this:

1. sex.bots are specifically written to appear human. In all cases they are programmed to aggressively deny the fact that they are bots. This makes them unique, as in most other cases I’ve researched, the machine makes it’s identity clear. The end result is that, at least initially, the unwary chatter will interact with these bots thinking that they are human.
2. I propose that a linguistic analysis of these bot’s scripts will reveal a lot about how gender roles are represented online. These bots are scripted in such a way in order to make them desirable to chat with. So they have to occupy optimal “fantasy” gender roles. It should also be noted that they potentially effect what chatters have come/will come to expect from chat partners. Looking to the future, these “fake” people may help dictate online behavior. They’re scripts also will help reflect how they bots creator’s view gender behavior as well.
3. They’re actively being used to make money. Sex.bots are online prostitutes/madams². And that’s a novel and scary application. For the most part other bots are productivity tools.
4. The rather, um, intimate nature of the human/bot interaction in these cases can be used to tease out issues about technology relations and interaction anxiety that have been theorized about for years (we can trace it at least as far back as Mary Shelly’s Frankenstein). And these are issues that will be projected into the future as well.
5. The semiotics of cybersex are rather rich. Cybersex allows for a unique interaction state involving the co-creation of metafantasies³. It’s a great subject to analyze. And it’ll be interesting to demonstrate how well this communication genre really is susceptible to bots for a variety of reasons.

And, frankly, no one is writing on this subject from this angle. And that type of academic novelty is a good thing. Would I rather be working on my longer term cellphone camera project? Frankly, yes. But that is a GIANT issue and the ethnographic recruitment for it would have sucked. So this is far more manageable and I can run pretty far with it.

^{1 – So far I’ve seen “porn-bot” used a few times, same thing with “chatter-bot” but that can cover a wide range of chat roles.}

2 – I’m considering the 24×7 ChatWalkers as a possible thesis title.

3 – I’ll get to the meta-fantasy thing in a future post.

The short reason is that the issue of human/AI interaction will be a huge field in the future. Its a matter of when, not if, we reach the point where computers will be able to consistently mimic human behavior in such a way that one would never know that they are interacting with a machine.

Trust me. I know.

How, you ask? Well, I’ve been there and done that.

During my webcam research I got suckered by a bot. It was my fault really. Here’s the basics:

I had been spending a lot of time researching Yahoo Chat rooms. My goal at the time was to help build data to disprove a notion: webcam’s primary purpose was sex/cybersex. There is a folks notion that either:

only exhibitionists had webcams
or
webcams created exhibitionists

This seemed like a gross simplification. But thinking it wasn’t enough. I needed to prove it (I’ll publish the proof theory at a later date). And proof required research. So I headed to the field: YahooChat rooms. And that’s when I started to observe bots.

For the most part they were easy to pick out. Very basic discursive scripts.

Then it happened: I had been hoping to get an interview with a cam whore (sorry mom… Really it’s a performative genre/category of webcam user). And I thought I found one who was open to discussion. And I thought I negotiated a deal for an interview in return for joining her site for a day. So far so good. But it turned out the site was a scam. And when I went to ask her what the deal was, I realized to my horror I was talking to a bot^{1 2}.

And that was that… I was hooked. And as I began to see large holes in my webcam direction, it became clear that I was barking up the wrong tree. So using an idea from the oblique strategies (which is another subject I’ll write on) I decide to kill my idea to save my thesis. And I moved on to bots.

^{1 – if you are interested in the entire painful experience, I’ve archived the entire thing. You can check it out here. IMPORTANT NOTE: As this interaction was held with a non-human participant, this does not need to comply with Institutional Review Board (IRB) procedure. As such the user name in this document does not need to be protected.}

2 – I’ve got a theory in the works as to how I (and other like me were fooled). I’ll get that posted in a few days.