But its fundamental proposition is implausible. Seen from a socio-biological perspective, and from a male point of view, Brooks presents the problem of sex as being historically how to acquire it, how to capitalise on it and how to ensure a ready supply. He envisages new technologies, and especially learning AIs, might overcome the inequitable distribution of sexual resources.

This is treating intimacy as a quasi-economic problem and omits the psychology that drives it, an omission that comes from a line of overoptimistic forecasts about the future which have explored the proposition that technology can reduce the complexity of human intimacy to the more manageable prospect of stimulation and button-pushing. Along the way, Artificial Intimacy uses the titillating pleasures of futuristic sex aids and lifelike dolls as "clickbait" to get you reading.

Along the way, the book contains some interesting portraits of the direction that technology is taking, from sex robots to algorithmic virtual friends anticipating our every need. Brooks divides the field into a taxonomy of digital lovers, matchmakers and virtual friends. He imagines the great danger of the technologies to be that of commercialising or exploiting our need and desire, seen in monopolies like Facebook and Instagram. And obviously, this has already been accomplished - these companies have been obscenely enriched by the manipulation of ordinary human desire for connection.

The book conflates various technologies and so the claims for their radical effects are over-stated. The matchmaking technology is no more "artificial:, although scaled-up, than the traditional marriage broker who facilitated matches within community norms. Likewise, Facebook & Instagram don't create a radically different effect of communication from the old-fashioned letter. Meanwhile catfish and scammers find a new route to an old end: fraud.

The funny thing is that while Facebook makes money surreptitiously from the ardent postings of friends and family, it really doesn't disturb the flow of feeling between real people. It creates an environment, it becomes a medium, albeit a distorted commercial one. It doesn't change the mode of human affect in the dramatic way that the book fears - it just cashes in on it.

The book works from assumptions of evolutionary biology, glancing across historical oversimplifications about sexual subordination and male privilege, from Bonobos to feudalism, that has a rhetorical effect similar to the algorithms under discussion - arousing interest but failing to satisfy.

The real danger of artificial intelligence is somehow out of vision in this book, and in similar portraits of the potential of AI in the field of intimacy. It imagines that gratification is the point of intimacy, which underestimates its psychology. It is of course quite possible to treat the lover as a sex object and to subjugate them, even to the point of annihilating their response. But this is a perversion of the place of intimacy in most people's psychological make up and it simply doesn't end up gratifying at all.

In championing the technologies of artificial learning, the book comes back to the dimension of sex as a problem of supply and demand. It sees sex, which is "intimacy" only in the most prosaic sense, as a good to be distributed or monopolised by power. This is a central project in socio-biology and the book is true to the evolutionary aspect that drives this field. It leads however to an unsatisfying analysis of sexual relations, leaving out the part where our feelings enter.

In fact, the central moral dilemma of love and intimacy - its strange oscillation between subject and object as found in the one person - is erased by these portraits of technological futures. Resorting to the imaginings of Hollywood and Netflix, the conclusion discusses The Handmaid's Tale, Westworld and Her as "the future in four fictions".

I think the most telling indicator of the book's failure to bring into focus the challenge presented by new technological sociabilities, is the book's admiration for the scenario put forward in Her. That film is a fictional portrait of a man's quest for simple emotional landscapes following his divorce. But his AI partner does not meet him at the place where he is. The fictional arc of the movie follows the classic "what-if" of artificial intelligence: what if robots became so advanced they were "just like us"? It is precisely at this point in the story that we see a wishful element arise - the robot becomes herself desirous. In the real world, this is not what artificial learning has taught us. It has shown that high-functioning simulation can be highly effective - and deceptive. It's not really intimacy at all.