Stu N
Regrettably, despite significant strides in establishing respectability over the past decade, the adult industry can still sometimes be a breeding ground for deceit, where unscrupulous individuals seize upon any new technological developments to exploit the industry for their own cynical gains. These opportunistic actors perceive the sex industry as an enticing opportunity for quick financial returns, undermining its integrity.
AI, crypto, and outrageous marketing claims: all of these seem easier to exploit in our industries than others.
The willingness of many in the adult industry to succumb to the seductions of new tech – literally any new tech at all – without doing any due diligence makes the industry vulnerable to the pursuit of newness at the expense of quality. That can be harmful: we’re still repairing the damage done by “revolutionary” sextech brand Lora DiCarlo, for example.
It’s wise to be suspicious—even cynical—of the effects of technology like artificial intelligence, especially in places where AI intersects with sex.
AI is moving fast, and it’s already butting up against political obstacles as society at large grapples with the advance of a technology that seems to have the potential to outpace us. Right now though, our understanding of AI is clear enough to peek into the very near future and speculate on how artificial intelligence might affect us in the adult industry tomorrow, rather than a century from now.
VR & AR: More real than real life
Perhaps the most immediate application for artificial intelligence is in virtual and augmented reality. These three dominions—AI, AR, and VR—seem to overlap conceptually, and it’s difficult to find any authoritative writing on AI that doesn’t mention VR too. Looking toward the future, they are two sides of the same coin. Apple’s recent announcement of the augmented reality headset Vision Pro demonstrates Big Tech’s conviction that this space is commercially viable.
Virtual reality has the hypothetical capacity to create almost overwhelmingly immersive sexual experiences that blur the line between what’s real and what’s digital, the effect potentially being more pleasurable, or at least more visceral, than the sum of its parts. VR porn, though, hasn’t yet found its feet, and so far, it’s been underwhelming. Whenever I’ve tried it, I found myself looking around the set instead of focusing on the action, trying to read the titles of the books on the shelf and examining the space for clues about the production, unable to suspend my disbelief.
Today, since porn VR isn’t really AI-based in any significant way yet (though SexLikeReal is working on it), it’s something of a curiosity more than a sexual encounter. It’s sexy only in moments and usually by accident, and it’s very difficult to resist actively seeking out the seams and flaws rather than suspending your disbelief.
It’s a question of budget. On one end, it doesn’t cost much to create porn that can be viewed in VR goggles, and it’s probably within reach already for most full-time adult content creators.
At the other budgetary end, full VR experiences are multi-million-dollar affairs. Half-Life: Alyx, commonly cited as the best VR experience yet developed, cost in excess of $50 million, and was designed to make a loss since the audience for it was relatively small. Personalized sexual VR experiences are some way off, without the backing of billionaire investors.
AI is already used in products like the SyncBot, which uses neural networks to match the movements of the toy with on-screen action. Of course, interactive content such as this isn’t really anything new. The AutoBlow AI+ also has an AI angle, having been trained on thousands of hours of explicit videos to ensure it gives a different experience each time.
Regardless, AI is the technology that will underpin the inevitable sexualization of virtual and augmented reality. First, though, a wealthy market for it needs to emerge.
Content and the means of production
For a while, it seems likely that adult content producers are safe from the risk of being replaced or made obsolete by artificial intelligence. AI-generated porn, for example, on the whole looks artificially generated, and most people who seek out fantasy porn are already well-catered to by the vast, vast 2D and 3D pornography industry, like Hentai and comic books.
Digital avatars are similarly unlikely to displace real performers and models: it’s not exactly the same thing, but digital popstar Hatsune Miku launched in 2007 and yet real people continue to make music. Hatsune Miku is a success of marketing, not necessarily of technology. That’ll never change.
What’s more, people simply enjoy making adult content. It doesn’t really matter if AI can do it better, many do it for the fun of doing it. Average people still play tennis despite the fact that Serena Williams is better at it than them.
In fact, it has become my belief that the adult space is about to become the most secure space for human creativity. Open AI models like ChatGPT are now under increasing political pressure to be cleaner than clean, safer than safe, and the result is a bizarre kind of prudishness in which ChatGPT will actually chastise you if you ask it a question of a sexual nature.
I got banned from ChatGPT in the course of writing this article because I kept asking it about buttplugs.
As a result of the political pressure, OpenAI models are voraciously self-censoring in an attempt to be seen as a “healthy” technology, and the models are incredibly resistant to talking about sexual subjects. ChatGPT, then, is probably very useful for the creation of content around things like doorknobs and hairdryers, but for those of us whose business is sex, we’ll just have to roll up our sleeves and do the work ourselves. I find myself unexpectedly reassured by this realization, despite the inevitability of the technology being continually ‘jailbroken’.
Getting up close and personal
AI’s impact on content creation will likely be behind the scenes. For example, AI neural algorithms can be used to dictate what kind of content should be produced by analyzing user data and collating detailed user profiles, from demographics and behavior to fantasies and aspirations.
Then, by comparing that data to others with similar interests, the system can identify patterns and suggest adult content that perfectly aligns with that user’s preferences. That’s called ‘collaborative filtering,’ and it’s already been widespread for some time – for example, in your Netflix suggestions.
In other words, while it’s unlikely the production of adult content will be affected in the near future, the content itself can be much more easily personalized and customized. Not ‘custom porn’ in the sense that a customer can pay a performer to do what they want to see, but that AI can be used to suggest relevant content in real-time according to the user’s intimate preferences.
These are incremental changes, small enhancements to already quite sophisticated search and categorization algorithms, designed to serve adult content to users in diminishingly improved ways. The major advances from AI are going to be made under the industry’s hood, in places like ethics, and in legal and regulatory realms.
Artificially intelligent convolutional neural networks are already used in the production of deepfake porn content, and AI porn generators exist by the thousands. I’ve already seen a handful of Japanese porn producers leaning on AI-generated porn to bulk out their content, and it’s impossible to know what datasets they’re training these models.
At the rather extreme end, the Japanese brand Milky Cat has been posting AI bukkake images in recent months, leading some of their followers to question whether they’d stopped producing real content entirely.
The laws around deepfakes and AI porn have not yet solidified – we don’t even enforce real revenge porn laws properly, so don’t expect any decent legislation to protect identities and intellectual property rights in porn within the next decade.
Poor data in, poor product out
Finally, there are the changes to the adult industry’s workforce and infrastructure that might be brought about by AI, changes which might seem unconnected to the production of adult content at first glance, but on closer inspection are likely to have deep and lasting consequences.
Roles that don’t have any inherent sexuality about them, like payroll and HR, can more easily be integrated with AI than, say, performing in front of a camera. But if you start replacing HR directors with algorithms, the old biases that already exist in hiring – for example, a preference towards straight white men in positions of power – become only more deeply entrenched.
This is already happening. Larger adult brands are already hiring third-party recruitment agencies which rely on AI to filter job applicants. But since the datasets on which these algorithms are making decisions are themselves rife with real-world prejudices, those prejudices are only made more efficient in practice. In other words: in an industry like ours that desperately requires better representation of marginalized voices, AI is still selecting straight white CEOs… only faster.
The same is potentially true in content editing and curation. AI tools already exist that can process video and edit it, and this is an extremely attractive proposition across the industry, from amateur content creators on OnlyFans to international studios alike. But editing is an art, handing that over to an algorithm will almost certainly lead to the homogenization of adult content.
There has long existed a phrase in the film industry to describe how poor quality input results in poor quality output: ‘SISO’ or ‘shit in, shit out’. It’s traditionally used to describe how poor footage will result in a poor finished product. But it’s equally apt when talking about AI algorithms: poor data in, results in a poor final product out.
Perhaps, in the longer run, that might be a benefit. We will see producers whose content feels different, feels human, rise to the top against an upswell of uniformly boring adult content.
Stripping it down: a final analysis
When we discuss AI and sex, there’s an instant, almost irrepressible temptation to immediately imagine legions of sex robots with seductive voices cooing and oohing and ahhing over an owners’ every sexual whim. We imagine AI-generated porn performers with that distinctive, uncanny sheen of inhumanity, performing sex acts of which our own, real, meaty bodies are anatomically incapable.
Those changes get the headlines. We tend to imagine the macro before the micro, but the truth is, AI isn’t going to drastically change the production of adult content immediately. Changes will be small, incremental, and behind the scenes.
