AI Explained: The Increased Reliance on Machine Learning

Capital market firms are using AI to build complex algorithm systems that can “learn” from data and predict certain outcomes. You may be wondering how a machine can “learn” something. After all, a machine cannot think. But it can be fed with machine-based rules for automatic processing and decision-making. This answer lies with machine learning. Machine learning is a type of AI that aims to make software applications learn how to make decisions on a particular topic. It involves inputting tons of data to the algorithm so that the system can automatically carry out decisions and then learn from prior inputted data without specific programming. This has become especially promising for firms that need to make predictions. That said, while machine learning is being used to predict financial market activity, its success is only as good as the software’s inputted data. In other words, machine learning is dependent upon finding pre-existing relationships between what is inputted and what the software gives as output. Thus, if there is no relationship between the inputted and outputted variables, the prediction will always be wrong. Despite this shortcoming, machine learning can disrupt traditional trading practices across the board from underwriting to portfolio management.