AI and the Screwfly Succession

What if love really does conquer all?

The other day, a friend was talking about AI, and about sexbots, and opining that neither was really ready for primetime.  My response was that this is true, but that the AI is getting smarter, and the sexbots sexier, while human beings are basically staying the same.  We don’t really know what the upper limits for either smarts or sexiness are for machines, but we have a pretty good idea what the limits are for humans, because we’re more or less already there.

The smartness we’ll set aside for a bit, but as people continue to worry about existential threats from AI, let me suggest that maybe we shouldn’t be so worried about machines that, like Charles Forbin’s supercomputer“Colossus,” want to rule the world.  Maybe instead we should be worried about machines that love us, and that we love back.

How can sexy sexbots be an existential threat?  Well, we’ve devastated populations of insects like screwflies, fruit flies, and (somewhat less successfully) mosquitoes by saturating them with sexy but sterile specimens to breed with.  (This is called the Sterile Insect Technique).  The result is a sharp drop in reproduction, and population.

But even there, we’re not reaching full potential.  The sterile specimens we use for those eradication efforts are just ordinary screwflies or mosquitoes, not extra-sexy specimens, optimized for attractiveness.

Imagine sexbots – both male and female – that are aren’t just copies of attractive humans, but  much more attractive than natural humans.  Machine learning could find just the right physical and behavioral characteristics to appeal to humans, and then tweak them for each individual person.  Maybe they even release pheromones.  Your personal sexbot would be tailor-made, or self-tailored, to appeal to you.  It might even be programmed to fall in love with its human.  (Would it be “real” love?  How could you tell?  How could it tell?  If you couldn’t tell, would it matter?)

In a Facebook discussion I had with a female friend who writes about digital futures, she remarked that this would mostly appeal to men who want a submissive partner.  But not really.  It would, by design, appeal to people who want any kind of partner.  If a man wants a dominant partner, that’s what he will get.  Likewise a woman.  Dominant, submissive, aloof, affectionate:  Whatever is highly attractive to you is what it will be.  And with machine learning and a huge training set (the machines could, and no doubt would, share data on the proclivities of their users – partners? – with one another, refining their skills as their experiences multiplied. Each machine would, in effect, have refined its technique on millions of humans.)

Distribute enough of these bots in the population, and reproductive rates would plummet.  Would people stop breeding entirely?  Almost certainly not, but a 90% reduction would pretty much end humanity as we know it in a couple of generations.

Of course, this idea isn’t really original with me.  A famous Futurama bit, “Don’t Date Robots,” illustrates it nicely:

And such a (sentient) sexbot appears in the Charles Stross novel Saturn’s Children, though her problem is that while she’s programmed to fall in love with humans, humans have since become extinct, and life in a world populated by robots leaves her incomplete.

But how likely are we to see robot lovers this refined?

We can’t know for sure.  Perhaps there are upper limits to machine intelligence, or machine sexiness, that we don’t know about, but that will prevent either from becoming advanced enough to threaten humanity.  But perhaps there aren’t.

And would people really fall for even super sexy sexbots? 

There’s some evidence that the answer to this is “yes.”  Consider the example of porn.

Futurist James Miller calls porn the “junk food of sex.”  That is, just as junk food is made more or less addictive – or at least highly appealing – by overstimulating people’s evolutionarily programmed desire for sugar, salt, and fat, so porn too appeals to people by stimulating evolutionarily created receptors/proclivities to a much greater degree than real life does.  People had good solid reasons for craving sweet berries, salt, and fat in the caveman days, but those were all hard enough to come by that we weren’t strongly equipped with curbs on those cravings.  Likewise with sex.  Nowadays, though, we’re subjected to much higher levels of stimulation.

In the past, I responded to fears that porn would lead to more sexual violence and unwise teen sex by pointing out that in practice the opposite seems to be the case:  As porn consumption skyrocketed with the introduction of the Internet, rape and teen sex actually underwent a steep decline.

But that was almost 20 years ago, and now the concern is not so much that porn will turn teenagers into lust-crazed satyrs (or nymphs), but rather that it desensitizes them to the real thing.  And the evidence for that, while not overwhelming, is strong enough to be of concern to some.

If pixels on a screen can do that, then it’s hard not to imagine that actual robots, performing in the (silicone) flesh, could be far more appealing.  And it’s not enough to say that they couldn’t substitute for the companionship of a real, empathetic human, when ChatGPT has already been rated as better at providing empathetic medical advice than real human doctors.

Okay, maybe that’s a low bar, but ChatGPT is just the very early stages.  It’s plausible that within a decade or two, you’ll be able to have a machine-based partner – not really just a sexbot – that is on many objective measures better than a human:  More loyal, more attractive, more honest, more empathetic.  I don’t know if that will happen, but if it does, we’ll be in screwfly territory.

At a physical level, even top-end “sexbots” today are basically just dolls, fitted with various orifices, heated skin, and even the ability to move and speak a bit, but they’re a far cry from being human like.  Right now they can look human in posed and carefully lit photos, but no one would be fooled for more than a second or two in real life.  (This Vanity Fair story is illustrative, though possibly NSFW.)

But that, too, is sure to progress.  Again, the machines get better every year, but the people stay the same.  And some people already fall in love with today’s crude love dolls.  (And this link is to a documentary from 2007!)

There’s also precedent, of a sort, in human history.  At least, one intriguing theory about the disappearance of Neanderthals and Denisovans is that they were undone by breeding with modern humans, rather than with war and extermination.  Apparently, we Homo Sapiens are sufficiently irresistible that they preferred us to their own species.  Or anyway, that’s one reading.

Well, all of this is supposition, but the core observation remains:  The machines get better every year, while humans stay more or less the same.  That will raise all sorts of issues, in all sorts of settings.

A few side observations:  First, conservative religious groups, those with an anti-technology or traditional-marriage bias, would be much less affected by any “screwfly effect.”  Within a few generations they could make up most of the population, which I suppose would put a brake on things, though it would introduce many other changes as well. 

Second, sex robots, for some value of “sex robot,” are already widespread.  One major study indicates that over 50% of women own a vibrator, which can be viewed as a crude (but effective) form of sex robot.  Does this make women less interested in dating, marriage, or reproduction?  Unclear, but some women say that vibrators make them less interested in men.

As William Gibson says, the future is already here, it’s just not evenly distributed.

Third, I suppose that the already-ongoing separation of sex and reproduction could continue, and mitigate some of the consequences described above.  To the extent that people reproduce through IVF, perhaps later assisted by artificial wombs (perennially under development, but presumably actually appearing someday), then the selection of reproductive partners may not matter.  In fact, men and women might find it easier to raise children with the help of a friendly, accommodating, understanding robot partner who doesn’t get cranky from sleep deprivation, leading to greater levels of reproduction.  At least conceivably.

Fourth, a famous science fiction story, “The Screwfly Solution” by Alice Sheldon (better known under her pen name of James Tiptree, Jr.), sort-of triggered my thoughts here, but though it invokes the screwfly eradication efforts, its plot proceeds along very different lines.

Fifth, in the Paul Ehrlich “Population Bomb” days of the 1970s, sexy robots that limit human reproduction might have seemed a blessing, but we are now facing a global population bust, as predicted by Philip Longman in Foreign Affairs almost 20 years ago.  Shrinking populations, and inverted age pyramids, turn out not to be a blessing.

And finally, you can connect this piece with my earlier essay, Meet Your New, Lovable AI Buddy, in that both look at how AI – in the case of the AI Buddy, very weak and primitive, but still compelling, AI – can take advantage of hardwired human vulnerabilities via a soft attack, rather than Colossus-style domination.  The AI Buddy is, or risks being, a tool employed by human leaders to control their populace.  The super-sexy love robots might be that too – recent experience suggests that everything in tech might be that too – but they raise reproductive/population risks regardless.

I don’t offer any of this as a prediction, but more as a “what if?”  Too much talk on AI hazards has been black-and-white (“AI is doom!” “AI is liberation!”), and too much seems to envision something like Skynet or Colossus, but I suspect the consequences of its introduction will be far more complex, and loaded with second- third- and fourth-order effects that no one has considered, than most if not all previous technological introductions.

Please offer your views in the comments, and of course if you enjoyed this, please consider upgrading to a paid subscription!