In the first of this series of two articles we looked at the impact of new technology on financial services.  I wrote that technology, particularly digital technology, almost always starts as a way to do something you already do more efficiently than you already do it.

A spear is more efficient than a rock for hunting deer. In the same way, a simple computerised ledger with manual input is more efficient than a man in fingerless gloves writing in a big book with a quill pen. And blockchain is way more efficient than a basic single-site computerised ledger.

Greater efficiency by necessity means that fewer people are needed to achieve the same output.  One hunter with a spear is more efficient at hunting deer than ten with rocks.  So nine hunters are now out of a job and sit in the cave complaining. Or so technophobes have been saying since (probably) the Neolithic.

But, it’s not long after adoption that smarter users begin to realise that technology can actually enable us to do new things – things we couldn’t do with the old technology.  

With spears our nine unemployed hunters are now able to hunt elephants – something they couldn’t do with rocks.   The tribe has more food, grows and becomes larger. And this is the promise of technophiles.  They say: “technology has always brought dislocation but then new jobs emerge and the economy rebalances – often at a higher level of productivity.”

We’ve seen these dislocations in recent years in media, publishing, leisure, retail, transport and more.  There are many fewer travel agents than 20 years ago – but an infinite number more social media coordinators.

The common theme across most recent dislocations of employment (recent being since the agricultural revolution) has been the automation of routine manual tasks by machine.  From seed drilling, through weaving to highly complex manufacturing, human performance of repetitive manual tasks has been gradually replaced by machine performance.

The fintech revolution is underpinned by AI, big data and cloud computing. They enable, increasingly, the performance of routine cognitive tasks by ‘machine’ or, more correctly, by software algorithms run on machines.   And the machines are programmed to learn – this doesn’t mean they are sentient, just that they can use massive data sets from billions of repetitions to refine their algorithms still further.

And yet, in areas where human interface is required they still fall heavily short of maximal efficiency.  Google and Facebook, despite billions of investment, still manage to serve up ads which are, at best, only passably relevant. They’re marginally better than previous models of advertising placement – but they are by no means perfect.

Returning to finance and fintech. It is highly likely that many of the internal functions of financial institutions currently performed by humans will be run by algorithmic technology in the future. Credit decisions, security, inter-institutional transfers, securities trading, asset allocation and many many more aspects of finance will be impacted.  After all – these functions are all about moving information from one place to another.  And digital technology suits that.

So, core financial functions won’t provide much new employment in the future. However, a whole raft of new roles is and will continue to emerge.  In management consultant speak there are in non-routine, cognitive tasks.  Tasks at which machines are poor and humans are excellent (or much better at least).  Simply put these tasks can be broken down into:  what should the machines be doing, are they doing it and how do our customers feel about it?

Business and policy leaders are now very much alive to the unintended outcome of large-scale technological interactions. Fake news bots are now a recognised threat.  There have also been well-documented examples of algorithms in financial institutions perpetuating discrimination against certain demographic groups in the granting of credit.  These malgorithms have produced highly negative outcomes not intended by their programmers.  

Humans are needed to define clearly the beneficial outcomes beyond efficiency that algorithms are to produce.  They are needed to monitor the data to ensure that those outcomes are being produced and to make changes when they’re not.  They are needed to provide an ethical dimension to the instructions given to machines.

But beyond such high-concept needs there is also a real need for humans to be expert at assisting machines to interact with other humans. To ensure that messy and irrational humanity gives the right data to AIs and that the output of that AI is suitable for humans. Humans are needed to sell to, to advise and to guide other humans through the forest of data and analysis being provided by machines.  They are needed to ensure that what the machines are serving to their customers is what the customers want.

Maybe, one day, bankers and financiers will become robots. All capital will be allocated efficiently and all risks managed without interaction of humans. But that day is a long way away.

Before then, those who are developing truly digital financial ecosystems need to be training people in governance, oversight, monitoring, sales, advisory and customer service for the digital marketplace.  The financial market of the future must have humanity at its heart. It must serve customers beyond machines and a purpose beyond pure profit.  It can only do that if guided by people.

The GCC Financial Forum 2018 - co-hosted by Euromoney Conferences and the Economic Development Board of Bahrain - will be held in Manama on February 20 and 21.
 

About the author

Richard Banks is consulting editor to Euromoney Conferences. The opinions voiced in this article are his alone.