Thursday 7 August 2014

Here Come The Robots?

Criticism of Silicon Valley's blueprint for a future society has begun to gather some momentum over the last couple of years. Privacy of communications is still perhaps the biggest source of public concern, along with cybercrime and the security of online shopping and banking, but questions about automation and its effect upon employment are now starting to be raised too. This is a concern that resurfaces periodically with every major wave of technical innovation.

In the 1960s the prospect of the abolition of work by automated machinery could still be treated as a utopian goal by groups like the Situationist International: back then trade unions were sufficiently strong that it was taken for granted that wages could be maintained as working hours shrank. The neoliberal reversal of the last 30 years has ensured that labour lacks any such power nowadays (if it ever had it). The matter cropped up again in the early 1980s when it looked as though the personal computer "revolution" might do away with millions of white collar jobs, but that turned out to be a false alarm too. In fact PC operating systems and application programs were so primitive and unreliable that many new jobs had to be created in IT departments tasked with trying to keep them all running.

The latest version of this problem is being raised just now, thanks to dramatic advances in robots controlled by AI ("artificial intelligence") software. Google's driverless car is one uncanny example, and the giant online retailer Amazon has been rumbling about employing pilotless aerial drones to deliver ordered goods to customers. It seems extremely unlikely that the powers who control airspace will permit this any time soon, but Amazon more realistically talks about AI-driven automation of the location and retrieval of inventory in its chain of huge warehouses, which poses a genuine threat to jobs that are already scandalously underpaid. In a recent interview with the online magazine Slate, Professor Andrew McAfee, a research scientist at MIT's Center for Digital Business, was asked by interviewer Niall Firth "Are robots really taking our jobs?" and he replied by offering these three alternative scenarios:
  1. Robots will take away jobs in the short term, but more will be created and a new equilibrium reached, as after the first Industrial Revolution
  2. Robots will replace more and more professions and massive retraining will be essential to keep up employment
  3. The sci-fi-horror scenario in which robots can perform almost all jobs and "you just won't need a lot of labour" 
McAfee believes that we'll see scenario three in his lifetime.

When asked further about any possible upside to this automation process, McAfee described the "bounty" he saw arising as a greater variety of stuff of higher quality at lower prices, and most importantly "you don't need money to buy access to Instagram, Facebook or Wikipedia". One doesn't need to have actually read Keynes to recognise that though McAfee might know a lot about robotics, his grasp of political economy is rather weaker. If employers "just won't need a lot of labour" then they just won't need to pay a lot of wages either, unless forced to do so by some agency whose identity is very far from obvious right now. If no-one outside that fraction of a percent of the population who own the robots has money to spend on food or housing, then the prospect of free access to Instagram and Facebook is unlikely to appease them very much. It's entirely possible that they will employ their spiffy new 3D printers to reconstruct Madame Guillotine, and Prof McAfee might perhaps be misremembered as a 21st-century Marie Antoinette for that line.

This blindness to the political - perhaps the most important victory the neoliberal ascendancy has achieved - is amplified a thousandfold in a survey conducted at the start of 2014 by the US Elon University and Pew Internet Project, in which 1,896 highly-qualified practitioners in the fields of AI, robotics and networks were asked to comment on this question of job loss. One of the survey questions asked respondents to share their answer to the following query:
"Self-driving cars, intelligent digital agents that can act for you, and robots are advancing rapidly. Will networked, automated, artificial intelligence (AI) applications and robotic devices have displaced more jobs than they have created by 2025? Describe your expectation about the degree to which robots, digital agents, and AI tools will have disrupted white collar and blue collar jobs by 2025 and the social consequences emerging from that."  
Respondents were fairly evenly split between three scenarios similar to those that McAfee proposed, and I was fairly unsurprised by the lack of any mention of real politics by any of them. I searched the summary of the survey results, to discover only a single occurrence of the word "politics". To be sure there were 20 occurrences of the word "political", but most of those instances conformed to a similar, vague template, something like:
"...our political and economic institutions are not prepared to handle..."
"...economic, political, and social concerns will prevent the widespread displacement of jobs..."
"...humans are in control of the political, social, and economic systems that will ultimately determine..."
"...unemployment should be addressed primarily by creating a smarter political system that serves the citizenry..."
These really are little more than pieties: mustn't appear too technologically deterministic, ought to mention social effects, there... done. Among this stratum of techno-utopians actual politics is regarded as something rather old-fashioned that happened before social networking, a type of natural disaster that only re-emerges at times of social breakdown (and some of them are of course actively engaged in deploying the new technologies to suppress dissent under those circumstances). If McAfee's third scenario were to come about it would certainly generate a faster rise in inequality even than at present - even than that envisaged by Thomas Piketty - and it would be likely to precipitate some sort of social breakdown. The question is, what sort of breakdown?

If the notorious 1%, the rentier class and their heirs, did end up with more or less all the money and all the property, what kind of economic model could they operate? Not even the most pony-tailed of tech-utopians believes that robots will be able to design themselves by 2025, and so a tech-elite will still be required to do that job. Here Slavoj Zizek's notion of the "surplus wage" comes in handy once again: the owning class can pay very generous salaries to those people who invent the robots for them, and those who work on their fabrication. Such a surplus wage, one not directly related to productivity, can always be withdrawn to suppress dissent, making membership of the tech-elite into a sort of lifeboat, with a queue of people waiting for your seat if you should stumble. (That implies that some degree of technical education must be retained to keep the queue full).

This wouldn't be an entirely unprecedented state of affairs as something quite like it already prevails in the popular entertainment business - movies, TV, music - at all levels below that handful of top stars who can extract enough to set up as producers themselves. Under such a model any resurrection of a labour movement and trade union power becomes all but impossible, since people who have neither jobs nor workplaces can't easily unionise, even if there were a will in the Labour or Democratic parties to reform anti-union legislation (which there isn't). For the same reasons any revival of Leninist/Bolshevik communism is improbable - no workplaces to organise in - while anarchist/mutualist movements like Occupy similarly lack any purchase on the real economy, as well as any adequate source of funding.

The most likely scenario would be emergence of some kind of new Jacobins, renegade members of the privileged tech-elite who stir up and manipulate mobs of the unemployed to attack the rentier elite. The US Tea Party already displays many characteristics of such a movement (it would need to turn against its Koch brother backers, but such about-faces aren't uncommon in the history of right-wing extremism). Putin's FSB-oligarch state has some of the right stuff too. In China such renegade factions already pose a threat to the ruling party, as recent purges of Bo Xilai and Zhou Yongkang suggest. A succession of such revolts would install more or less identically impotent juntas, who might persecute and expropriate the super-rich for their own gain but fail entirely to restore employment (a bit like Argentina then). Fairly quickly the capacity to conduct advanced electronics research would be eroded and the robot age would grind to a rusty halt.

Personally I doubt that McAfee is right about his third scenario, not because large corporations lack the will to throw most of us out of work (they do not) but because the abilities of AI have always been hyped way beyond the reality, in order to extract grants from ignorant and gullible politicians. Everyone forgets that the fighting drones which the USA wields to such devastating effect in Afghanistan and Pakistan are controlled by people, not by AI computers. The Russians have just announced an autonomous war robot, a small armoured car on caterpillar tracks equipped with a radar-, camera- and laser-controlled 12.7mm heavy machine gun. It's being deployed to guard missile sites and will open fire if it sees someone it doesn't like the look of. Now there's an IT department I wouldn't want to work in...

Nevertheless such speculations are far from useless. Like the climate change debate, they concentrate minds on what a short time window we have to prevent such horrible future outcomes. Preserving incomes at the cost of profits is a matter for politics, and for unfashionable class politics at that. We  need to be inventing and researching a swathe of new policies, from John Lewis-style mutual ownership, through universal basic incomes or negative taxes to job shares and reduced working weeks, that might gain electoral appeal if McAfee's second scenario turns out to be the more likely.

GILT BY ASSOCIATION

I don’t have any special credentials as a commentator on geopolitics, but occasionally, like now, I feel obliged to have a stab at it. The c...