The Age of Em review – the horrific future when robots rule the Earth

In the future, or so some people think, it will become possible to upload your consciousness into a computer. Software emulations of human brains – ems, for short – will then take over the economy and world. This sort of thing happens quite a lot in science fiction, but The Age of Em is a fanatically serious attempt, by an economist and scholar at Oxford’s Future of Humanity Institute, to use economic and social science to forecast in fine detail how this world (if it is even possible) will actually work. The future it portrays is very strange and, in the end, quite horrific for everyone involved.

It is an eschatological vision worthy of Hieronymus Bosch. Trillions of ems live in tall, liquid-cooled skyscrapers in extremely hot cities. Most of them are “very able focused workaholics”, who “respect and trust each other more” than we do.

Some ems will have robotic bodies; others will just live in virtual reality all the time. (Ems who are office workers won’t need bodies.) Some ems will run a thousand times faster than human brains, so having a subjective experience of much-expanded time. (Their bodies will need to be very small: “At this scale, an industry-era city population of a million kilo-ems could fit in an ordinary bottle.”) Others might run very slowly, to save money. Ems will congregate in related “clans” and use “decision markets” to make important commercial and political choices. Ems will work nearly all the time but choose to remember an existence that is nearly all leisure. Some ems will be “open-source lovers”; all will be markedly more religious and also swear more often. The em economy will double every month, and competition will drive nearly all wages down to subsistence levels. Surveillance will be total. Fun, huh?

This hellish cyberworld is quite cool to think about in a dystopian Matrixy way, although the book is much drier than fiction. Hanson says it reads more like an encyclopedia. But if it’s an encyclopedia, what are its sources? The physicist Niels Bohr was quoting an earlier Danish wit when he said: “Prediction is very difficult, especially about the future.” But Hanson’s book is, in part, a defence of prediction. “Today,” he complains, “we take far more effort to study the past than the future, even though we can’t change the past.” Yes, you might respond: that is because we literally cannot “study” the future – because either it doesn’t exist or (in the block-universe model of time) it does exist but is completely inaccessible to us. Given that, the book’s confidence in its own brilliantly weird extrapolations is both impressive and quite peculiar. Hanson describes his approach as that of “using basic social theory, in addition to common sense and trend projection, to forecast future societies”. The casual use of “common sense” there should, as always, ring alarm bells. And a lot of the book’s sense is arguably quite uncommon. The governing tone is strikingly misanthropic, despairing of current humans’ “maladaptation” to the environment (the low birth rate in rich countries, and our excessive consumption of TV and even music apparently prove this), and there is an unargued assumption throughout that social patterns and institutions are more likely to revert to pre-industrial norms in the future.

The major difficulty in the analysis, though, lies with Hanson’s vision of how ems will think of copies of themselves. If an em decides to terminate itself and have a saved copy of an earlier brain-state reawakened, is that archived version still the same person? Will a briefly lived “spur” copy of an em be happy to be terminated after it finishes the task it was created for? Hanson assumes there is no big problem about the continuity of identity among such copies, and therefore erects a large edifice of sociological speculation on how the liberal use of em copies and backups will change attitudes to sex, law, death and pretty much everything else.

But there is plausibly a show-stopping problem here. If someone announces they will upload my consciousness into a robot and then destroy my existing body, I will take this as a threat of murder. The robot running an exact copy of my consciousness won’t actually be “me”. (Such issues are richly analysed in the philosophical literature stemming from Derek Parfit’s thought experiments about teleportation and the like in the 1980s.) So ems – the first of whom are, by definition, going to have minds identical to those of humans – may very well exhibit the same kind of reaction, in which case a lot of Hanson’s more thrillingly bizarre social developments will not happen. But then, the rather underwhelming upshot of this project is that fast-living and super-clever ems will probably crack the problem of proper AI – actual intelligent machines – within a year or so of ordinary human time. And then the age of em will be over and the Singularity will be upon us, and what comes next is anyone’s guess.

What about, you know, us? Early on, Hanson cheerfully says: “This book mostly ignores humans.” If meat people survive in the em era, he says, they will probably live far from the cities on low pensions. Given that this future is so gloomy for just about everyone, one does end up wondering why Hanson wants to wake up in it – he reveals in the book that he has arranged to be cryogenically frozen on his death. I suppose it is at least possible that, one day, he could open his eyes and have the last laugh, as he surveys the appalling future he foresaw so long ago.