Finding the exact porn that scratches the moment’s whimsical itch can be a real drag: Part of it is just figuring out what you’re in the mood for — but you still have to cruise the digital buffet of flesh until you hit the jackpot, and most of the time, you’re probably settling. Dr. Taylor Kohut, a psychologist at the University of Western Ontario who has researched porn for a decade, thinks he can do better. Kohut is currently crowdfunding a site and app called the Porn Genome Project, where participants map images of porn, effectively metatagging sample pornographic images to compile data he says will help us better understand arousal and improve sexual content.
“It can be frustrating clicking around to find what you’re looking for,” Kohut told me of the typical porn search. “Part of it has to do with the way we sort and categorize media, which is very coarse and crude,” he said. “The categories are broad and nonspecific, and the algorithm relies on too few subtleties. We believe that if you increase the distinctions that can be made, you have a better chance at matching things that are more similar to a person’s actual interest.”
Categories like MILF, tentacle porn, lesbian, bisexual and gay are perfectly adequate starting points, Kohut argues, but from there, existing porn is often not searchable to the level of detail a user would prefer. How does one, for example, quickly locate porn where the actors appear to be genuinely, believably turned on by each other? What’s the keyword for that?
“People always respond more positively to porn when it seems like the people are authentically enjoying themselves or communicating pleasure or at least good at lying,” he explains. “But that hasn’t been singled out by the marketplace or researchers as an important variable. Even beyond that, we don’t understand much about the specificity of sexual preferences. It’s an area of ongoing active research, but when we think about it, we adopt crude typology like am I gay, straight, bisexual? Do I like vanilla sex or am I kinky?”
Let’s say, for instance, that Kohut is in the mood to watch anal sex. “I’m shown all my choices, but the way that’s depicted can vary dramatically,” he said. “The characteristics of the performers, their age, height, weight, eye color. Whether it’s in the context of a dyad, a gangbang, a threesome, whether or not there’s any explicit plot. Or what motivation there is for doing it — what’s the relationship between the characters?
“Maybe I specifically want a redheaded man to be plowed by a pizza delivery boy who’s been working out at the gym. That’s kind of difficult to search by keyword, and even if I could find that one video, the next related video suggested will have one or two of those properties, but it won’t have that degree of specificity.”
Then consider how much time he’s now spent just looking around. “I might even not be interested in masturbating anymore at that point,” he said.
What’s needed, Kohut argues, is more data to aid in the diddling. The Porn Genome Project is seeking $50,000 on IndieGogo to build a website that will be loaded with public domain pornographic imagery ripe for the mapping. Participants will be volunteers who only need a brain, eyes and ears, and preferably a willingness to look at pornographic images and identify very specific things about those images.
“With your unoccupied hand, you will indicate whether an image contains a naked woman, a bed, a pair of fishnet stockings, or any of the thousands of other genes and indicate where in the image each of these features are located,” the description on the site reads. “Once you’re satisfied with your work, you can move on to a new image, and ultimately sequence as many additional images as you’d like.”
Take the example Kohut put on the IndieGogo site to illustrate the issue:
“Generally, people see the image on the right, with the red masking, as more pornographic than image on the left,” the caption reads. “This is intuitive because you can’t see her clothes, but is also counter intuitive when you take into account that there is actually significantly less skin exposed. None of the methods that are currently available for classifying pornography properly account for both the presence and the absence of visual cues that people use when making such judgments (presence of exposed skin AND the absence of a swimsuit).”
While this sounds like an incredible level of tedium, it’s the stuff of better data, and therefore better porn, and ostensibly, better orgasms. “Rather than patronizingly assuming that a preference for blondes means all videos with blonde women are right for you, our techniques will allow us to build detailed individualized profiles of exactly what elements you want and don’t want to see in your porn,” the site explains.
Sussing out these kinds of distinctions isn’t just about making sure everyone who watches porn experiences a customized level of maximum titillation. It’s also about understanding the connections between images and arousal, but also images and lack of arousal. A porn user might go looking for some lighthearted spanking online and think they’ve found it, only to stumble onto something considerably more violent in a video than they bargained for, and be turned off by it.
Parsing when porn is violent or not is another aim of the project. Kohut says there’s a perception based on at least one study that something like 88 percent of porn is violent, even though other research found as little as 2 percent. “It’s all over the place,” Kohut says. “If we want to establish that seriously, we have to get on the same page about the definition of violence in pornography. The techniques we are trying to develop will help us get a better sense of how specific characteristics group or don’t group together to help us come up with a naturalistic definition of violence in porn.”
Kohut turned to crowdfunding because he’s up against some challenges when it comes to getting funding. In his experience, he says it’s difficult to get funding for any research that portrays porn positively. Utah’s governor just declared porn a public health crisis, calling it “evil, degrading, addictive and harmful,” he notes. And reams of research have sought to find links between porn and damaging effects, arguing that it leads to sexual dysfunction.
“Porn is associated with promotion of unsafe sexual practices, like condomless sex as well as other forms of sexuality,” Kohut nots. “It all gets roped in there, and the framing these days is not moralism, but public health,” he said.
Yet Kohut’s own research contradicts some of these perceptions. He has found that porn use does not correlate with sexual aggression and that porn users are less sexist toward women. And his soon-to-be published study found porn use in relationships is typically a positive thing, assuming both people in the relationship use it. The Porn Genome Project is not just about improving porn itself, but also about improving porn research.
“What we want to do is generate some information that can help us have a sober second look at porn, what things co-occur and what things don’t,” he said. “And it will be useful for informing laypersons, media and researchers, all groups that tend to be loaded with assumptions about the nature of porn.”
Kohut will make his data available to anyone interested in seeing it, and if the Porn Genome Project is funded — the deadline is Tuesday, Jan. 17 — maybe soon we can all find what we’re looking for.
Tracy Moore
Tracy Moore is a staff writer at MEL. She covers all the soft sciences like psychology, sex, relationships and parenting, but since this is a men’s magazine, occasionally the hard ones. Formerly at Jezebel.
