Artificial Intelligence: will trading robots soon replace human traders? |

Artificial Intelligence: will trading robots soon replace human traders?

In recent years an important phenomenon has been at work, in the finance world as a whole, and in particular on the finance markets. And as the days tick by, it is getting bigger and bigger. We are talking about the automation and use of computer programs in the trading rooms of banks and other financial institutions.

This is a companion discussion topic for the original entry at

I think you’ve touched on the ‘elephant in the room’ here.

Not something people in the financial industry want to admit for several reasons, not least being, as in most other industries where it is also happening – job security.

Artificial intelligence (AI), robots, automatons or computers – whatever you want to call them have been ‘taking over’ many aspects of trading for years. As is easily recognised in the case of stock trading algorithms which do the majority of buying and selling. Most equity transactions now actually take place in a data centre far away from the noise and humanity of the old trading floors.

There are various explanations why we can still see human faces milling about on some trading floors round the world given that computers dominate stock trading everywhere.

Few of them stack up under close scrutiny as most insiders readily agree that if you were to start from scratch today, trading would be fully automated. So on the surface it seems the only reason humans are still around is as a form of historic showmanship.

However, when you dig a little deeper there are other seemingly valid reasons for maintaining the human element.

The anti-AI brigade readily point out the numerous examples existing of over-reliance on robots and insist that the human touch is important during a crisis.

One prime example used to illustrate their point is the Knight Capital ‘disaster’ in August 2012 when they lost over $450 million in just thirty minutes due to a software error.

Another example has been termed the Flash Crash of 2010. When at the New York Stock Exchange (NYSE) in May 2010 the Dow Jones Index (DJI) dipped by around 10% and then miraculously rose again – all in just a matter of minutes. This caused the likes of Procter and Gamble share value to fall to next to nothing whilst Apple shares surged 40,000%. Subsequent investigations revealed that robotic trading was the root cause.

Others point at the stock market crash of October 1987 when computerised trading programmes, the forerunners of AI as we now know them, were held to be largely responsible.

So maybe they do have a point here – especially when you consider that computers execute trades in milliseconds.

As explained by Cambridge University neuroscientist John Coates, who was previously a Wall Street trader:

"a return to human-only trading is hardly viable given the difference in profits robots can generate in comparison to human traders.

The fastest human cognitive processing takes around 200 to 300 milliseconds according to Coates, while a high frequency trading (HFT) can process at around a millionth of a second.

Achieving hundreds of million trades a day, robot traders’ minor profit margins soon add up, and they dwarf those achieved by humans."

So this transactional speed can be seen to be partly blessing and partly curse.

Being programmed with complex algorithms robots tends to identify trading patterns to buy or sell and once a strategy is undertaken their sheer momentum can only exacerbate volatility in the market.

This and their inability (as yet) to process and react to external share price influences and announcements in the way a human trader should means there is the likelihood that they will, in all probability, incur significant losses in such situations.

The other associated concern with AI trading is the lack of transparency continuing to plague the HFT industry as the speed at which computers operate makes tracking individual trades nigh-on impossible without a full audit trail.

Incidents of robot traders ‘manipulating’ the market are on the increase – spoofing, or dynamic layering, has emerged as what some see as an unfortunate but inevitable exploitation of HFT by a minority of less trustworthy trading firms and individuals.

In an ideal world being able to combine the natural human intuition with the processing capacity of computer systems would be the way forward combined in a cybernetic organism interface - think Startrek here - which I am sure is being developed somewhere as I write this.