We recently learned that the Pentagon was testing an artificial intelligence capable of predicting future events several days in advance.
GIDE gathers data collected by satellites, radars, submarines, cyber and intelligence elements, and shares it in a cloud. All this information is then available to a Pentagon machine learning and artificial intelligence system that can quickly analyze and process it.
Algorithms could analyze the average number of cars in a parking lot in enemy locations or identify how many planes are parked on a runway. If it detects a change, the AI could issue a warning. This would allow the Pentagon to act proactively, rather than waiting for episodes to occur, according to the commander.
Most of us don’t aspire to that much and would settle for a small AI that would simply tell us if the stock market was going to go up tomorrow or not. That would be enough to make us millionaires. It doesn’t seem so difficult either, does it?
How does AI work?
In theory, if a trained expert can predict something from data, AI should be able to do the same. All AI models basically do the same thing: they find patterns in the data, and then use those patterns to predict what will happen next.
AI relies on data, so you need data to train an AI. But not just any data: you need clean data, in a particular format. In some cases, you have to label and categorize the data first.
So, to train an AI model that predicts the stock market, all you have to do is feed it a bunch of well-organized historical stock market data. Sounds easy, right?
What’s the problem with AI stock market prediction?
The first problem is possibly that randomness can overcome expert stock market predictions. Even if that is not true, the second problem is that the stock market is a level-two chaotic system. Unlike level one chaotic systems, such as the weather, stock market predictions themselves can change the stock market.
If someone develops an accurate AI stock market prediction system, they can use it to make money for themselves in the stock market, or they can sell access or information. However, if accurate AI stock market prediction were to become commonplace, it would influence the stock market in unpredictable ways. This would mean that the patterns it learned from historical data would no longer apply, making its predictions less accurate.
Does stock market prediction work with AI?
The answer is “it depends” and “sometimes”. For obvious reasons, researchers and investment firms devote a lot of effort to stock market prediction with AI. However, the results are mixed.
Some of the best-known AI-managed funds are I Know First and World Markets, while Toggle, AIStockFinder, Danel Capital, BetterTrader and Finbrain offer AI-generated investment advice. Trade Ideas targets beginners and focuses on teaching them how to trade.
However, one must weigh the possibility of losing money. K1, an AI-driven hedge fund, cost one investor more than $20 million, while Aidiya, another AI-driven hedge fund, closed after achieving only a 2% return.
AI can predict the stock market. However, like any investment, it carries risks, and past performance is no indication of future results. Money can be lost and managed funds can fail. For the time being I’ll stick with Warren Buffet’s human intelligence or with this hamster!