Stock trading scams are hardly new. What is fairly new, though, is the hype around AI trading platforms. Rising public awareness of AI trading bots, allied to promises of infallible profits, has galvanized scammers and turned some obvious truths and truisms into a toxic cocktail.

Let’s take a step back and consider the problem. Experienced traders either can, or like to think they can, identify repeating patterns in the price movements of stocks. Knowing these patterns, they believe, gives them an edge, an advantage that their competitors lack.

Show even an absolute beginner a historical stock chart, taken over most time periods, and they will be able to point to repeating patterns of ups and downs in the movement of any stock. Sometimes these patterns look as predictable as the pattern on a heart monitor, sometimes not.

However (and here’s the rub): you can see the pattern plainly once it is laid out, but what about that blank space on the page margin to the right of the chart? That space represents the future – and it is blank because it hasn’t yet happened.

Ask that same newbie to draw their best guess for how the price chart will move in the empty space, and patterns get so much harder to predict. Then ask them how much money they would like to bet on their guestimate of the future being precisely accurate.

What about adding mathematical analysis to the problem? It has been done for decades. It is called quantitative investing. The whole idea here is to devise mathematical algorithms to sift huge quantities of data in order to produce useable investment strategies.

AI and machine learning have taken this further and have opened up new possibilities for discovering meaningful data patterns in vast data sets. This is great for funds that still have humans looking at the proposals for stock picking and stock selling that the AI and machine learning bots spit out. You can find any number of investment funds these days that advertise both their use of AI to find investment opportunities and the fact that they still have human fund managers in charge of the overall process.

AI investing comes in yet another guise, this time as AI tools or toolsets designed to be used by DIY investors. This term refers to individuals who set out to play the stock, currency or commodities markets on their own, by signing up to a trading platform.

From there, individual investors formulate their own investment strategies, along with the entry and exit points for anything they invest in.

Many of these trading plat- forms have their own inbuilt AI-driven tools, or else are amenable to the use of third- party AI-driven toolsets. It is worth pointing out here that the Financial Times recently noted that more than 70 per cent of DIY investors lose money, with or without the help of AI tools. Another investment news site, howtomoney, puts the figure of those who lose money even higher, at around 90 per cent. What this should suggest to any sensible person is that it is worth having a good, hard think and doing some soul searching before plunging into DIY investing, with or without the benefit of AI tools. What are your grounds for thinking that you will be the one in every ten DIY investors that will actually generate positive returns?

So, what do these AI tools, which are used by individual and professional investors alike, actually do? There are several different answers. Some are de- signed to, or claim to, pick stocks or currencies or cryptocurren- cies that are ripe for buying.

Others focus more on looking at socially determined reasons for entering a stock purchase. Here, the AI analyses vast amounts of news and online social media in real time to iden- tify key events that can move stocks.

An example of this is Lev- elfields, which has what it calls ‘pioneering AI and language processing technology’. In the company’s own words: “Lever- aging decades of experience in machine learning and contex- tual analysis, we created a novel Al platform to identify patterns from market-altering events so that we can all quickly find these events and understand how to react to them.”

The technology, Levelfields says, screens millions of signals related to more than 6,000 stocks, sifting through the noise to find those events that affect prices, and using historical data to expose patterns and put those events into context, so stock and options traders can better un- derstand entry and exit points.

In other words, it is still up to the DIY investor to decide whether or not to act on the buy or sell information that they are being given.

The holy grail, however, for a DIY retail investor would be a pure, AI-driven platform where the AI decides what to do and executes all the trades. All the investor has to do is invest a basic sum then sit back and pocket the profits on a monthly basis. The machine does it all and dumps the profits into your account. Wouldn’t we all love to have a money tree like this?

The problem with this con- cept, of course, is that markets require people on either side of a transaction.

If a wonderfully efficient AI trading system appeared and was embraced by everyone, we’d all be buying the same thing at the same time and selling the same thing at the same time.

This raises the question: who would be playing the role of the idiot on the other side of the transaction? How would they stay solvent enough to keep being on the losing end of every trade?

It is worth remembering that stock returns come from earn- ings distributed to investors as dividends, plus the market’s ‘feeling’ about a company’s likely future growth and earn- ings performance. Profits are not magical. Without productivity growth where are increased

There is already a huge scam in progress in Australia, where an anonymously run AI plat- form, Quantum AI, is touting its ability to return significant prof- its on a minimum investment
of AU$400 (£205 ). Disturbingly, this claim is being promoted via deep-fake videos of well-known personalities ostensibly vouch- ing for the system’s effectiveness.

The adage that if something looks too good to be true, it probably is, holds true here as much as it does anywhere else.