Why you should expect less from AI algorithms in finance

The said algorithm would predict exact product prices in the future, disrupt financial markets as we know them, and make regular trading obsolete.

Regardless of whether it’s actually going to happen, this “waiting mode” amongst investors actually restrains them from taking advantage of what AI has to offer today – like solving other, tangible problems in trading.

You cannot have too much of a good thing – or can you?

The number of ETFs (exchange-traded funds) has grown six-fold since 2007. Today, one can invest in almost 7000 different products of this kind. Simultaneously, in the last 10 years the number of mutual funds worldwide grew by 40,000. On top of that, financial institutions compete in offering their clients the widest product universe possible. Can you see the problem?

These days, anyone who prefers to choose products on their own is overwhelmed with the choice. It distracts investors and affects their ROIs. Instead of looking for new opportunities, we are stuck analysing which product is the right one. We have dozens of options on how to invest in one industry, and the choice of the right product has become a problem of its own.

It's all relative

Imagine two cars in a car park, just about to start racing. One of them is fast – can you guess which one? Sure, you can make assumptions before the race, but we all know it depends on what is under the hood. Moreover, “fast” is relative. It would be much easier to make that guess while the cars are already in motion, right?

The same goes for racing to get the highest ROI – it’s about guessing which product is the best, but that race is already on. Sounds simpler, and from the perspective of AI algorithms’ predictions, it really is.

Finding the best product to invest in may not be about getting the best product anytime and anywhere. It may be about choosing the best one out of a comparable group. It is something that existing AI algorithms can cope with and it is within investors’ reach today. It would solve the growing problem of product choosing, and bring substantial value to the people.

Algorithms like that have plenty of input data already available – because that race has been going on for years. Product quotes, economy indicators, previous market drop scenarios – it is all there for algorithms to learn about, assess and adjust to.

AI can be wrong 49% of the times

Just like in blackjack, when you have the right strategy, what earns you money is the number of deals you have. In investing, the proper trade management can bring you solid returns, even if the AI algorithm you use is not 100% right.

As Peter Lynch said years ago: “In this business, if you’re good, you’re right six times out of ten. You’re never going to be right nine times out of ten”.

That is why it may be a sound option to put some more trust in AI, even if it is not always right with the predictions. Instead of waiting for the Holy Grail, some consistent outcomes should be enough to start investing with AI and make profits using the newest technologies.

What are we waiting for?

As investors, there is a good chance we got lazy and waited too long for a tool to predict the future. If so, maybe we should take a step back and look around to see whether AI algorithms can have a softer form. Like a super-indicator, which can relatively tell you which product is better. And when you have that, the race for profits may look completely different for you.