David Levy’s Love & Sex With Robots had an unexpected effect on me; I discovered that I’m a romantic after all.

No, wait—it’s not what you think! I haven’t suddenly developed an abiding adoration of androids as a result of Levy’s points. Quite the opposite: I’ve always believed that I have a very rational and cynical view of human love and affection, but Levy’s deployment of the psychological and physiological foundations of those emotions in support of his thesis has made it clear that I have, to some extent, been fooling myself.

Levy’s thesis is that not only will it be possible for humans to be doing the wild thing with fully artificial partners within around five decades or so (possibly less), but that the majority of us will inevitably be doing so, with little or no sense of shame. The first part of the conjecture is more than plausible, especially if you’re plugged in to the RSS feeds coming out of the electronics, computing and physics establishments. The Japanese, especially, are making massive strides in robotic appearance and capability—aided, perhaps, by their very different cultural attitudes towards automata—but as Levy points out there’s no shortage of effort in the US and elsewhere, albeit from different angles. The possibility of robots that could pass for human in a casual momentary encounter is far from being a wild speculation. Frankly, I’d say that fifty years is probably a conservative estimate of our ETA on the far side of The Uncanny Valley.

What is a little tougher to accept is the notion that, during said casual encounter, you’ll decide that you want to have sex with this friendly and attractive android. I’m willing to bet you can think of three or four major objections without even having to make an effort; I know I did. Levy has heard them all, though, and plenty more, and he proceeds to refute them all. And this is where my hidden romanticism reared its atrophied yet leonine head: for the great majority of Levy’s refutations are based on statistics and psychology, both of which are disciplines with a deserved record for, more frequently than not, discovering exactly the results their practitioners set out to find.

So, when Levy quotes studies wherein a significant number of responders used anthropomorphic language when talking about their computers or toy robot dogs, I find myself wondering what options they had to choose from on the questionnaire. When he quotes surveys wherein people self-report on the frequency of their liaisons with human sex-workers, I find myself remembering all the times I’ve gleefully lied through my teeth to the person who has cornered me with a clipboard while I try to buy some new shoes. Psychology has value, but it isn’t predicated on measurable variables; I don’t trust it like the “hard” sciences, which it masquerades as here.

That aside, Levy’s points aren’t actually that incredible when taken in isolation: we do develop affections for non-human (even non-humanoid) things; we are prone to getting our sexual jollies by whatever means necessary (given our particular circumstances and cultural barriers). But these aren’t the bits that really stuck in my throat. I found it harder to deal with the notion that artificial intelligences will develop well enough to not only simulate the aspects of a human personality that could make someone fall in love with it, but also to reciprocate that love.

Now, if you’re even slightly more romantic than me—and there are pieces of Brutalist architecture which meet that criterion—you just had a huge emotional knee-jerk reaction. How could a robot love you back, even if you fell in love with it? And this is where I just can’t bring myself to square with Levy’s ideas: he claims that expert systems will be able to accurately mimic the reactions and behaviours of a sentient consciousness falling in love, and then uses the “quacks like a duck” argument. In other words, if it appears to be falling in love with us, might we not just as well accept it as being the same thing?

Logically, there’s a lot to be said for Levy’s arguments, but love and logic feel about as far removed from one another as Earth and the moon, and I can’t help but feel there would be something strangely exploitative about creating beings whose only purpose is to make us feel happy and loved, and to satiate our every sexual whim. Perhaps he’s right, though; indeed, perhaps learning to fall in love with artificial beings will encourage us to take a more rational view of love as a physiological and psychological phenomenon in years to come. Maybe I’m less a romantic than I am a cynic, though, because I don’t intend to hold my breath. I’d also argue that economics is going to provide a stumbling block; how many people will be able to afford a robot companion? Where’s the energy to power them going to come from, in a world where we’re not entirely sure how we’ll supply basic amenities to the entire global population fifty years hence?

My disagreements aside, Levy’s book brings together an interesting cross-section of research material about human-human and human-machine relationships, and as such provides plenty of food for thought for the sf-nal thinker, even though many of the throwaway cautions and “what ifs” in his conclusion have been well trodden long beforehand by genre writers. I’d very much like to give him Chris Beckett’s The Holy Machine to read, providing as it does the dark flip-side to Levy’s somewhat wide-eyed Church Of The Soft Machine best-case-scenario. And therein lies the fault, I think; an sf reader will find Love & Sex With Robots to be lacking the rigorous interrogation of ideas that the better fiction writers can provide. The happy corollary being, of course, that there are a hundred stories waiting to be written in response to it …