Can artificial intelligence replace human traders?

AI can streamline trading processes, but the human touch is still in demand for investment decisions.

Artificial intelligence is at the peak of popularity, making waves in nearly every industry and continuing to grow quickly. By 2030, the global size of this market is expected to surpass $1.8tn (see chart). In the stockmarket, the hype around AI has led many to speculate that machines will soon dominate trading, but these expectations may be going overboard.

Despite advancements in large language models (LLMs), we are still far from Artificial General Intelligence (AGI) — a true AI that can reason like a human. This level of technology is at least another five to seven years away, and until it comes around, human intelligence will remain a vital component in trading that won't be easily replaced.

While AI's capabilities are often overstated, it is capable of handling various tasks just fine. For example, major banks JP Morgan and Capital One are using AI-powered chatbots to streamline customer services. While some clients are not fully comfortable interacting with bots, the technology is helping reduce burdens of human employees and cut operational costs.

This suggests the main avenue of AI usage in financial companies will focus on creating call centres without ‘live’ employees, combining LLMs and social engineering to bring down customer support costs.

In trading, AI has also begun to play a major role in analytics, processing vast amounts of market data, news reports, corporate events and other financial information. In the past, sorting through raw data required much effort from human analysts. However, modern AI systems, like the BloombergGPT chatbot introduced last year, can better manage the workload, replacing junior analysts and reducing need for human hands at this stage.

However, that’s not enough to make real people obsolete. Strategic thinking, creativity and intuition all drive successful trading, and for now, all these remain firmly within human expertise.

Fast or smart?

A big debate among traders choosing strategies is whether to prioritise speed or smartness. For high-frequency trading (HFT), speed is everything, though, artificial intelligence isn't really required to achieve this.

Success here often depends on quality of hardware, investing in servers close to exchange data centres, which minimise delays and allow faster trade execution. Given the high level of competition in this environment, the speed at which data is processed plays a critical role, down to a millisecond.

Even in long-term trading, where strategy plays a bigger role, AI’s current capabilities are limited. Human traders still have the upper hand when it comes to deeper understanding of market trends and ability to make informed predictions. Likewise, when we need to test investment strategies and determine if they are still viable, there are simpler, time-tested methods that work just fine, including back testing and analysing historical data. The latest AI innovations are hardly necessary here.

Traders did try to explore use of AI systems several years back. They leveraged deep learning networks to analyse microstructure of the order book and predict its behaviour over relatively longer (for classic HFT) time frames, ranging from hundreds of milliseconds to one to two seconds. But as this technology became more widespread, market inefficiencies went down, and opportunities to profit decreased.

Trading is a battlefield of rival parties constantly waging war on each other over greater profits, which is why intuition and creativity are so important. After all, what could be more creative than battling for our financial future and well-being?

AI systems may help streamline tasks and process data, but creativity is precisely what machines cannot yet take away from us humans. That’s why final decision-making that truly drives success will remain in human hands for years to come.

Alexey Afanassievskiy, executive director and head of portfolio management, Mind Money