My son (who’s 6 and knows about these things) says that I’m too dumb to be a droid. It’s true that I may soon be put out to grass, replaced by an infinity of smart machines. But for now it’s me writing this piece and not a computer.

And what I want to say isn’t all that encouraging for the droids.

Unlike many, I generally welcome the advance of robots. Indeed I'm frustrated they've taken so long to move from sci-fi imagination to daily reality.

Replacing dull, monotonous work is good. Most of what robots do well is what humans do out of grim necessity, not love: easily definable, repetitive and precise tasks.

Against expectations though, the great majority of jobs still look pretty resistant to automation. In time robots may become good at everything from changing children's nappies to providing counselling, and I'd quite like there to be a robot that could clean my house.

But sadly that still looks a long way off.

Robots taking jobs?

That's a surprise. All my life robots have been about take our jobs. This was the message from Isaac Asimov in the 1950s, and hoards of both optimists (like Marvin Minsky) and doom-mongers (like Jeremy Rifkin) from the 1960s onwards.

A generation later the argument moved on, and since at least the 1980s it's been forecast that, having done away with millions of blue collar jobs, robots were now ready to demolish white collar jobs too. 

20 years ago, when I ran the think-tank Demos, I published a pamphlet on the imminent demise of lawyers, accountants and other professions who would be replaced by smart algorithms.

I'm afraid to say that that prediction sits alongside my forecasts of revolution in Saudi Arabia and mail order clones.

Most of the occupations that should have been wiped out have grown in numbers during this period. Oddly too, during the forty years since most futurologists predicted the end of jobs and the end of work, employment rates have tended to rise rather than fall (albeit with a significant shift away from men and towards women).

This is not to say that there hasn't been a lot of automation, from car factories to Amazon warehouses. But at a global scale, the big story of the last few decades was the shift of production to Asia and that was less a shift from labour to capital and machines, and more a shift from expensive to cheaper labour.

A similar picture

Today most futurologists sound identical to their forebears. They repeat the forecasts of a few decades ago warning that hundreds of millions of jobs will disappear.

The picture they paint (again, not unlike their predecessors) is one in which factories turn into sprawling complexes covered with robot swarms. Offices will hum with machines but no excited chatter at the coffee machines. Our cities will be dominated by driverless cars, and the frail elderly will remain at home waited on by robotic carers. 

Much of what robots do - sensors, smart limbs, analytic capacity and memory - will be disaggregated.

But a few decades later than promised, some machines will bring these capacities together, and look at least a bit more like the robots of public imagination.

As that happens we'll have to learn new rules and protocols - how to talk to your driverless car, your Servbot, your Carebot and your home (and as Jaron Lanier points out we may have to dumb ourselves down to use them well).

To that extent the technological determinists will be right - society will have to respond to the new technologies (for more on this see our upcoming Hot Topics event on Emotional Machines).

But history suggests that change happens in much more dialectical ways than futurology usually recognises.  Changes elicit other changes. Trends generate countertrends. New concentrations of power prompt coalitions and campaigns to weaken them.

As a result, simple linear forecasts in which technologies just wipe out jobs are pretty hopeless and misleading. This is partly a matter of economics. To the extent that robots do replace existing jobs, relative price effects will kick in.

Automation doesn't have to mean less jobs

Those sectors where productivity dramatically increases will see price reductions, and spending will shift over to other fields which are harder to automate: personal coaches, tour guides, teachers, carers, craft workers.

Their relative price will probably rise (as will that of highly skilled jobs in supervision - making sure the robots work, though these too will diminish over time). Labour markets have proven to be pretty dynamic over the last two centuries, coping with massive destruction of jobs and equally massive creation too.

There is no obvious reason why a much more automated society would necessarily have fewer jobs. Some very thoughtful economists are now arguing that 'this time is different', and that even if the warnings were wrong in the past, they're now right.

The nature of communications and information technologies means that they really will scythe through professional jobs. They may be right. But given how wrong comparable analyses have been in the past, we should start from a position of scepticism.

Valuing people

We also need to think dialectically about demand. Experience suggests that what we want in a more automated economy won't be the same at it is today. We may well be willing to spend a lot on truly smart robots to serve us, drive us or guide us.

But automation will also raise the status and desirability of what's not automated. Craft is booming in part because of robots. At the upper end, designer crafts fetch high prices for their imperfections as well as their perfections. Handmade is now desirable. So is hand grown.

These can now charge a premium where at an earlier stage of economic development they were seen as substandard. We should expect even more of a shift towards valuing people.

Face to face services are already a lot more expensive than commodities, yet at one time they were cheaper. But there is no sign whatsoever that demand for coaches, trainers, masseurs, beauticians is saturated.

Indeed, to the extent that automation further releases disposable income for other tasks it will shift the balance of the economy even further towards services, and in particular high touch ones.

People have brains

But the most important reason to think dialectically is politics. This is hard because the great majority of technologists and futurologists have no feel for politics (and often no feel for history). Implicitly they often assume that the public are dumb and passive - rolled over by big trends which they have no hope of influencing.

But two hundred years of technological revolution should have taught us that technological determinism is always misleading - mainly because people have brains, as well as interests.  

In practice people campaign, lobby, argue and organise. It may take them some time to get the measure of a new bunch of technologies. But before long they become agents rather than victims.

The ultimate capitalist dream

That's why the simpler stories of singularity are quaint fantasies but not much more. And it's also why the conspiracy view of the spread of robots as the ultimate capitalist dream - an economy with only consumers and no workers - is equally fanciful (though it is intriguing to ask why governments give such favourable tax treatment to investment in robots compared to employment of human labour). 

If no one was paid, no one would buy the products produced by robots. Henry Ford had to pay his workers enough to buy his cars. Similarly, a seriously automated economy has to work out some way of generating demand.

In theory every citizen could become a capitalist - and just enjoy a flow of dividends from robot companies which they then spent on consumer goods produced by robots. Or they could rely on handouts from the state. Alternatively rewards could be concentrated in the hands of a very few.

The point is that in any scenario, questions of distribution quickly come to the fore and open up very obviously political choices.

This highlights the broader point. Any new technology sets in motion political battles over who benefits and who loses.  The internal combustion engine prompted new rules and regulations like speed limits, and new kinds of provision like public buses.

Electricity prompted great utilities and public service guarantees as well as a huge apparatus of safety rules. Robots too will prompt similar discussions not just on regulation and law, but also on such issues as their use within public services or even the right to have one. 

The number of variables involved make it hard to predict where we'll end up on questions of ownership, privacy and provision. But I expect that the most likely result of heightened political argument will be that we demand not just that we have robots to serve us, but that we also demand to incorporate some robot strengths into ourselves. 

Join 'em

We may conclude, in other words, that if you can't beat 'em, join 'em. Indeed, if humans have any sense they will demand the best of what robots have: prosthetic limbs; synthetic eyes; expanded memories - so that they can keep the interesting jobs, and the status and pay that goes with them, rather than allowing these to be parcelled out (a topic covered at a recent Hot Topics event).

Since, as I pointed out earlier, robots are best thought of as disaggregated capabilities, we will surely want the best of these for ourselves, with our own brains doing the aggregation.

That's why a movement for human enhancement is more likely than any kind of singularity, and will make much of the current debate about jobs look anachronistic.

So when my son says I'm too dumb to be a droid, my answer is: yes, that's true for now. But I hope that I'm just smart enough to take the best of what the droid has, and that the droid is too dumb to stop me.