The robot banking machine doesn’t add up


Until the birth of true AI, robots are merely low-cost human proxies, a means to take cost out of large organizations and maximise economic benefits internally to fewer and fewer people.

A robot sitting at a deskNews that Bank of America’s Merrill Lynch financial advice arm is about to be joined by robots has set the internet chattering, not least because Bank of America has been the world’s largest wealth-management operation since it acquired Merrill Lynch in the epicenter of the 2008 financial crisis.

There are delicious ironies in the news, given that wealth – AKA other people’s debt – mismanagement precipitated the global crash that brought Merrill Lynch to the precipice of bankruptcy seven years ago. Can robots do any better than wreck the world financial system?

The robots in question are neither the cute NAO models favoured by Japan’s Mitsubishi UFJ bank, which earlier this year installed them in branches to greet customers; nor are they like the sundry humanoids serving in Nagasaki’s Henn-na robot hotel, which are there to slash HR costs and maximise profits and efficiency.

They – or rather, it – is a prototype automated investment platform. In 2016, it will be set to work on the sub-$250k accounts that are the preserve of the bank’s Merrill Edge wing.

The bank’s reasoning is not that the platform will outperform its human equivalents, necessarily, but that its very futuristic newness will help attract younger investors, in the same way that smartphone/app-only accounts delight customers who are too busy staring into their palms and making swiping motions to risk human interaction. One financial analyst said that Bank of America wants to prevent young investors from getting “hooked” on a rival platform, which makes it sound like crack cocaine.

In short, it’s a disruptive technology, and disruption is apparently what a financial sector that used to be about such boring things as stability, probity, and the tedious predictability of compound interest needs. At least, that’s the opinion of the bankers whose quest to minimise risk on a micro level seems to be accompanied by a thrill-seeking ramping-up of risk globally so that the ultimate beneficiaries, billionaire offshore investors, never get bored.

Lemmings vs. cliffs

The herd-like culture of modern banking was demonstrated when the instant response to the plans of Wells Fargo – which has the most human skin in the game – was that they were, according to CFO John Shrewsbury:

A real threat to our business, because we are disproportionately full-service, high-value-added, person-to-person activity, which isn’t for everybody.

Wells Fargo knows all about threat: parent company Morgan Stanley reportedly lost up to 80%  of its market value in 2008. We can infer from Shrewsbury’s comments that the sector’s response to automation must be to become less full-service, less value-added, and less focused on people, which would seem to be missing the point of the crisis.

But the big question is: will robots work in finance? In many ways, of course, it’s an absurd question to ask, because the core condition of any robot is that it can only follow instructions: algorithms, written by human beings, that produce pre-defined outcomes. The desired outcome of most investment advice is “make money”, which forces onto customers the same binary choice that faced the banks pre-2008:

  1. Boring –  make safer investments that, historically, have been shown to grow predictably over a long timescale (bearing in mind that Western economies are in recession one-third of the time). This option tends to favour old industries, such as oil, tobacco, energy, pharmaceuticals, and guns.
  2. Exciting – gamble, and hope that a more risky investment pays off spectacularly. This favours new technology, and buying the debt of people who can least afford to repay it. (Just ask Merrill Lynch, Morgan Stanley, or the population of Greece about the latter.)

We all know which option the banks chose.

The human machine

As I explained in a previous article on robotics and automation, banks’ talented, skilled, empathetic human employees are already functioning like robots, because the companies for which they work have become giant compliance systems. As anyone knows who has sought financial advice recently, or tried to open an account in a physical branch, you’re talked through a series of slides and tick boxes, the output of which is ‘Yes’ or ‘No’. That’s precisely how robots work, so those processes may as well be automated.

It stands to reason that the outcome of increased automation can only be as good or as bad as the thinking behind the algorithms. Put simply, they will either succeed or fail faster, cheaper, and more efficiently, which will at least lower the organisation’s costs – which is not the same as mitigating against risky behavior.

News of the innovation comes at an intriguing time: publication of a detailed report, also from Bank of America, on which disruptive technology is the most promising and the most likely to be profitable (for banks).

The company identifies three broad trends: the Internet of Things (IoT), what it calls “the sharing economy”, and online services. Robotics links all three, which to any bank taking the macro view means: “robots make money”.

The report says:

We have entered a period of accelerated innovation, made possible by the confluence of many complimentary enabling technologies required to change a business model. Think of it like this: big technology innovations often require many smaller technology advances first. This is the ‘building block’ approach to disruption, and we believe the majority of building blocks are now in place.

But what might the impact of robotics be on human employment, given that, in banking at least, human employees are becoming inconvenient cost centres in what is, quite literally, now a money-making machine? (One that, in order to make more money, needs to focus less and less on adding customer value, as we have seen.)

The report warns:

Robotics could worsen labor market inequality: the number of industrial robots is up 72 per cent in the past 10 years while the number of US manufacturing jobs is down 16 per cent… Technology poses major government policy issues with regard to income inequality, privacy, and cybersecurity.

Robot-1The report also states the bank’s belief that “tech” is a “micro-deflationary” force, but not “macro-deflationary”, and adds:

Innovation and the disruptive role of technology is a theme-rich cluster as the corporate and government world shifts from one of ‘too big to fail’ to one of ‘too big to succeed’ via e-commerce, e-security, robotics, and genomics. Today’s technological disruption is universal, and trends in data transportability, cloud wars, the generational shift away from ownership toward experience, wearable tech, and crowdfunding, will create substantial divergence in investment winners and losers in coming years.

Despite all this, the report claims that the historic argument for technology-driven unemployment is thin, and it uses the performance of the US economy as proof. In the US, technology disruption is both endemic and central to economic prosperity, and yet overall employment stands at 94.5%. This may be true, but the bank’s ‘building block’ analogy of a gradual accretion of discrete, disruptive advances may, conceivably, also mean that the same building blocks are in place for rising unemployment.

My take

The problem with this whole argument is that, over time and until the birth of true artificial intelligence, robots are merely low-cost human proxies, a means to take cost out of large organisations and, thereby, maximise the economic benefits internally to fewer and fewer people.

So why aren’t prices falling, so that people externally can benefit too? Technology itself is commoditising, of course. In other sectors – such as the media – content is being given away free, which has the effect of stuffing every channel with noisy intermediaries and advertising.

But prices elsewhere tend to be kept artificially high in order to maximise the profit potential inherent in reducing production costs, automation, and outsourcing to countries with low labour costs (where automation will also increase, to combat rising wages).

So the key thing missing in all such reports is this: while overall employment may not be falling – yet – and while many large organisations (such as banks) can increase their profits through automation, the ability of more and more human beings outside of these organisational megastructures to make a decent living is collapsing.

This forces more and more people to pick up ad hoc income from the breaking apart of other industries, such as by becoming an Uber driver, or by putting their apartment on AirBnB, or by offering their freelance services via a cloud platform (often for micropayment rates). This pushes down not only the end-user costs of service-based activities, it also reduces their perceived value and strips out regulatory protections.

Sooner or later, the taxman will clamp down on all those ad hoc, citizen-generated, peer-to-peer activities – just as he has on the thousands of small businesses that are based on eBay – while the robotic megacompanies continue to move their money around the world to avoid paying tax to national exchequers.

In the long run, this means that the only people who are paying for the maintenance of local or national services are the people who can least afford to do so. Inevitably, they’ll try to make money by offering services privately via global cloud platforms, playing cat and mouse with the local taxman.

Unfortunately, this also means that the industries that are most resistant to automation – all those definitively human activities that include writing, designing, making music, and so on – are now regarded (as I said in my recent report on Pandora) as low-paid support acts to an ever-growing, rapacious advertising industry. Makers are expected to provide their services for nothing.

Put differently, the money-making potential of activities that can’t be automated is falling, unless your clients are wealthy dilettantes on the one hand, or wealthy aficionados on the other.

So automation really isn’t giving us all the free time to be creative, imaginative people, sitting on mountainsides having Platonic thoughts about the nature of reality. There’s no money in it, and over time, there will be less and less. My advice? Learn robot maintenance and move to a tax haven. But whatever you do, don’t take up gambling.

In the meantime, governments must hold the banks to account and force them to pay for the damage they did to the global economy, and not punish those people who are least able to look after themselves locally. There really is no alternative.