Data and algorithms and artificial intelligence to manipulate them are all the headlines these days. A few weeks ago AI experts were highlighting some of the more extreme dangers of intentional misuse of these technologies. The large-scale hacking of the physical infrastructure that they warned of hasn’t – yet – come to pass, but the recent Facebook/Cambridge Analytica revelations are proof enough of what can be done without going to those extremes to be disquieting.
But just as concerning if much more insidious are the underlying changes to the economy brought about not by malicious intent but by the evolution of technology itself, with AI to the fore.
And here’s the thing. I am not a techno-determinist. I strongly believe that the course technology has taken has been shaped directly or indirectly by ideology, incentives and tax regimes, among other things, and if we put sufficient mind to it (a big if) we are capable of shaping it in other directions, towards other, more socially favourable outcomes, for example. That said, however, it is impossible to go back. We are where we are. And that is at a point, or so some people think, where today’s technology, formed by the influences mentioned above, is becoming self-organising. It is evolving semi-autonomously, outside conscious human control, and as it does so it is creating an invisible second ‘intelligent’ economy that is steadily subsuming parts of the physical one. Software is eating the world, in the famous phrase – including companies and jobs that will not reappear.
In two articles in the McKinsey Quarterly, 2011’s The Second Economy and Where is technology taking the economy? In 2017, Brian Arthur explains why. Arthur is a highly regarded complexity scientist and economics professor at the multidisciplinary Santa Fé Institute, and several years ago he wrote a highly original book about technology that I reviewed here. He said that technology is more like chemistry or biology than physics, building out from itself in ways that were non-linear and organic.
In the two essays he takes his insights further. Basically, he argues that just as steam power and then electricity bulked up the pre-modern economy by supplying muscle power that enabled mass-production and the huge increase in physical stuff that accompanied it, so the combination of computers, the internet and now cheap ubiquitous sensors is supercharging it by backing up the physical economy with a neural system – a virtual back office if you like – where more and more of the coordination, administration and linkages get done.
Perhaps more accurate than software eating the world is that it is modularising it, generating ‘libraries’ of digital modules available for use, with transformational effect, right across the old industrial landscape. Take driverless cars. It’s no accident that the frontrunners in autonomous vehicles are tech titans like Google and Apple, or start-ups Uber and Tesla, rather than traditional automakers, struggling to stay in the race. Data will be the most important component of the cars of the future (as long as they exist), not metal.
Modularisation is increasingly breaking down conventional industry sectors, blurring their boundaries and reshaping them into loose clusters of technology-linked suppliers, competitors and customers that behave more like ecologies than separate industries. In transport, some observers see the outlines emerging of a ‘mobility ecosystem’ based not on car ownership and mass production but on flexible individual preference – expressed perhaps in subscription models providing access to a variable combination of private, public and shared transport, refined as data from vehicles, the road and personal preference is collected and processed in the virtual economy. If the data and connections are valuable enough – for example generating a market for onboard information and entertainment – the transport element might eventually come free.
Even if it doesn’t, the direction of travel is clear. Increasing portions of the physical economy will migrate to the hidden digital one – and all industries will be affected. Arthur doesn’t mince words: ‘I think it may well be the biggest change ever in the economy. It is a deep qualitative change that is bringing intelligent, automatic response to the economy. There’s no upper limit to this, no place where it has to end’.
He is in no doubt that as with previous great economic shifts, many jobs will disappear – as they are already doing – but after agriculture and manufacturing, this time it is the last repository of employment, the service sector, that is in the firing line. It is correspondingly harder to see where new ones might come from. As Arthur notes, invoking another historical precedent: ‘Offshoring in the last few decades has eaten up physical jobs and whole industries, jobs that were not replaced. The current transfer of jobs from the physical to the virtual economy is a different sort of offshoring, not to a foreign country but to a virtual one. If we follow recent history we can’t assume these jobs will be replaced either’.
That of course poses a problem. Jobs don’t just provide work. Up to now, they have also been the vehicle for distributing wealth and ensuring access to the fruits of production (thank you, Henry Ford). But this benevolent circle is now breaking down under pressures that are partly ideological but increasingly technological, as networked developments feed on each other. The issue now, at least in the leading economies, is no longer production. In the US, for example, if total household income were shared among all households, the mean would be a enough for a decent middle-class living. Instead the agenda-topping item is distribution.
We are shifting, suggests Arthur, from the age of production to the age of distribution. Distribution being political rather than technical, like production, it is likely to be trickier to deal with – although Europe, with its experience of social legislation and safety nets and suspicion of free-market fundamentalism, may be in a better starting position to do it than the US. But there is no doubt that just as past dislocations needed far-reaching institutional innovation (pensions, the welfare state, trade unions) to palliate transitions and smooth rough edges, similar large-scale adjustments will be required this time.
Meanwhile, technology will continue to feed off itself as it grows like an invisible root system beneath the tangible economy. It’s a remarkable thought, if slightly creepy in its implications, given how effectively the unscrupulous have learned how to bend existing technologies to nefarious ends. ‘We just put information [like ‘crooked Hillary’] into the bloodstream of the internet and watch it grow,’ explained Cambridge Analytica CEO Alexander Nix before he was suspended. ‘Give it a little push now and again over time to watch it take shape… So this stuff inflitrates the online community with no branding, so it’s unattributable, untrackable.’ As well as economists, those anxious AI experts may just have landed themselves something else to worry about.