The recent boom around AI has led to a fully fledged change in market temperature, with mutual funds and retailers alike power-buying the large American tech stocks expected to benefit from AI – NVDA, AMD, MSFT and the likes.
Everybody is talking about using AI for generation of text, images or code. As exciting as the topic is and being hyped by the media, I believe people might be getting ahead of themselves. AI is certainly useful, and I have used it to write my own computer code for different applications. I’m certainly neither late adopter of novel technologies, nor a luddite – but the AI that exists and will exist for the foreseeable future is just not compatible with the hype and the images in people’s heads. See a great essay by Per Bylund on the topic here.
Similarly, when it comes to the market, I assert that AI will never fully replace human traders. While AI may excel in processing vast amounts of data with lightning speed, be trained to identify chart patterns and analyze company fundamentals, it lacks the essential human element that underlies successful trading & speculation – the ability to decipher and navigate the complexities of human psychology and emotion.
The stock market is a living, breathing entity that thrives on the ever-changing emotions, motivations, and reactions of its human participants. It is a battlefield of human psychology, where the slightest shift in sentiment can cause ripple effects. AI, no matter how advanced, cannot truly understand and process the nuances of human psychology and emotion that drive market movements – simply because it cannot experience them and their often sheer visceral intensity.
AI cannot ‘feel’ or ‘understand’ the fear that many long-term stock holders had during the 08/09 market crash, nor can it have ‘trust’ in an established trend even when technical sell triggers might appear. It cannot grasp the role of sentiment forming market bottoms – at least not in real time when it is most important. AI cannot understand market exuberance, animal spirits, nor can it tell whether a market is truly good or whether just everyone’s opinion of it is good.
What is the downfall of most human traders is also what sets them apart from machines, because humans can both learn to control their own impulsivity as well as exploit that of other market participants for a profit. Humans possess the unique ability to read between the lines, listen to or disregard propaganda, smell a rat, or discern the shifts in the market based on an emergent picture that sometimes only the subconscious can decipher, and which messages are often in stark opposition to what the financial news, classical financial or economics education, or other market participants say and do.
Intuitive understanding of human behavior is crucial in predicting market trends and making informed decisions. AI may be able to analyze data patterns, but it cannot intuitively grasp the emotional and psychological factors that are integral to successful trading. Human traders have the capacity to adapt to changing market conditions and make rapid decisions based on their instincts, experience, and the market barometer of their own and others’ sentiment. AI, while programmed to learn from past data, cannot match the innate ability of human traders to adapt and make adjustments on the fly, specifically when it comes to a psychological interpretation of ‘what’s happening’.
An “AI” devoid of both the ability to experience emotions and a long life history of having dealt with them, interpreted them, battled and overcoming (or succumbing to) them, is not able to make the right decisions, at least not at the right time. Most impactfully, humans possess the ability to learn from their own experiences and mistakes, which is an essential aspect of long-term success in the stock market. AI, on the other hand, can never evolve in the same way that human traders can – because this learning process is more than the mere sum of its parts of technical and fundamental anlysis. It is a highly contorted, iterative, abrasive and at times illogical process of putting cause and effect together by joining elements that often have no seeming connection at all. A process unfit for understanding by a purely logical entity.
Another big force driving the market is groupthink, the herd mentality. Many a trader has set out with clean intentions of being their own master, but their plans have fallen upon meeting the market, its participants, and their opinions and ‘expert assessments’. As Jesse Livermore said, “a man cannot be convinced against his own convictios – but opinions can talk him into a state of uncertainty and indecision”. And that can be overkill in the market. In the same line, AI and language models trained by big data off the internet would become a monstrous consensus mouthpiece almost by definition, following the herd of lemmings over the cliff after being mass-trained on their ideas, hopes, wishes and idiocies. The more AI enters the markets, the more decisions are made by the same blue-print – thus, the more volatility will leave from the markets and AI’s financial performance approach mediocrity.
The market exists as a viable vehicle for making profits just because the vast majority of its participants does not understand how it truly works, driven by groupthink.
Sure – ChatGPT and the likes are great tools, and I do not want to deride the technology at all. But everything has its place, and while it may offer some advantages in the world of stock trading, AI will never fully replace the unique skills and abilities that human traders possess. It can be never more than a tool to perform menial tasks, due to the above-mentioned shortcomings – which are not temporary but structural to AI.
