The City has to cut its costs and artificial intelligence will help it to do so.
AI is a hot subject and last week’s international summit at Bletchley Park in Buckinghamshire focused on the problems that AI may bring. But let’s step back from the fun, for alas Elon Musk’s idea that all jobs will be obsolete is bunk. Instead let’s think through the economic consequences of it all.
I’ll come back to the City in a moment. First, let’s accept that all new technologies bring problems as well as benefits, which is why we bring in regulations to reduce the harm. But that has always been so.
Look at motoring. In 1930, the year before the Highway Code was issued, more than 7,300 people died from accidents on British roads. Every death is a tragedy, but last year, despite there being vastly more cars on the roads, the number had fallen to 1,695. The world will develop a Highway Code for AI.
Next, while AI has been around since the 1960s, it wasn’t much use until the combination of huge amounts of data and huge increases in computing power over the past few years.
Universal problem: The harsh truth is that City services are far too expensive
Right now, it is very hard to predict the winners, or indeed the duds. But as argued in this column a month ago, I agree with former Prime Minister Gordon Brown that this is the technology that will enable service industries to boost both their productivity and the quality of their output. That means, contrary to current gloom, that the world will be able to increase both living standards and human welfare in the years ahead.
At a national level, this is good news for the UK. We are the second-largest exporter of services in the world, after the US. If, as seems to be happening, global trade shifts away from moving goods to selling services, this is doubly good news. So how do we make money out of it?
It is very hard to foresee how any new technology will develop commercial functions, and why some will take off and why others will prove blind alleys. However, we can make a decent guess at where to look.
Geographically, it’s the West Coast of America, where there are thousands of companies developing promising AI applications, many of which are attracting the interest of the giants. Look at the way Microsoft has made a huge investment in OpenAI, the most important of the pioneers.
Here in Britain, the best place to find a tiny enterprise that will end up worth a billion or more is in either Cambridge or London. As you might imagine, the hunt is on to find them. The trouble is that there are thousands of potential winners, but only a tiny number will strike gold. That’s wonderful if you can find one, but extraordinarily difficult even for experienced early-stage investors, and impossible for the rest of us.
Instead, let’s focus not on the developers of the technology but at the industries that apply it most effectively. The game will be partly to cut costs by substituting expensive people with relatively cheap technology, and partly to use some development of AI to provide a better service.
All industries that use highly-paid people will benefit: healthcare, education, business services, lawyers, and let’s all hope, the public sector too. The biggest winner of all, however, should be financial services.
But how? A huge amount of experimenting is going on right now, and already a few things are obvious. For example, drafting financial documents can probably be done just as well by AI as by a human, though clients would understandably be miffed if they were to pay for a known and trusted contact only to find that most of the work had been done by a robot.
Economic research is using AI to cut costs, though given the performance of our profession, I wish there was less focus on costs and more on judgment. That is where Musk is wrong. It is not about people being displaced by technology, but rather about how best to use technology to enable workers to provide a better service at a cheaper cost.
The harsh truth is that City services are far too expensive. It is a universal problem, but let’s focus on one area, fund management. Here in the UK, half a million people have a SIPP, a self-invested personal pension. But managing that typically costs 1.5 per cent to two per cent a year.
So if your real return on the portfolio is only four to five per cent, a third of the income is disappearing in fees. The prizes will go to the managers that can do as good a job, or better, on 0.5 per cent. At last they have a tool that may, if cleverly deployed, allow them to do it.