The following essay is reprinted with permission 
The Conversation is an online publication covering the latest research.
Artificial intelligence-powered tools like ChatGPT have the potential to revolutionize the efficiency, effectiveness, and speed of human work.
And this applies not only to healthcare, manufacturing, and nearly every other aspect of our lives, but to financial markets as well.
I have been studying financial markets and algorithmic trading for 14 years. AI has many benefits, but the increasing use of these technologies in financial markets also presents potential dangers. A look back at Wall Street’s past efforts to deploy computers and AI to speed up trading offers important lessons about what it means to use them for decision-making.
Program trading fuels Black Monday
In the early 1980s, boosted by technological advances and financial innovations such as derivatives, institutional investors began using computer programs to execute trades based on predefined rules and algorithms. This allowed us to complete large transactions quickly and efficiently.
At that time, these algorithms were relatively simple and were mainly used for so-called index arbitrage trading. This includes trying to profit from the difference between the price of a stock index such as the S&P 500 and the prices of the constituent stocks.
As technology has advanced and more data has become available, this type of programmatic trading has become increasingly sophisticated, allowing algorithms to analyze complex market data and execute trades based on a wide range of factors. rice field. These program traders continued to thrive on the unregulated trading highway, with over $1 trillion in assets traded daily, dramatically increasing market volatility.
Ultimately, this led to the massive stock market crash known as Black Monday in 1987. The Dow Jones Industrial Average suffered its biggest drop in history at the time, and the pain spread around the world.
In response, regulators have taken a number of measures to limit the availability of program trading, including circuit breakers that halt trading in the event of significant market volatility or other restrictions. However, despite these measures, program trading continued to grow in popularity in the years following the crash.
HFT: program trading with steroids
Fast forward 15 years to 2002, when the New York Stock Exchange introduced a fully automated trading system. As a result, program traders have given way to more sophisticated automation using more advanced technology, i.e. high frequency trading.
HFT uses computer programs to analyze market data and execute trades at very high speeds. Unlike program traders who buy and sell baskets of securities over time to take advantage of arbitrage opportunities (price differences in similar securities that can be exploited for profit), high-frequency traders rely on powerful computers and high-speed networks. Use it to analyze market data and execute trades at lightning speed. High-frequency traders can trade in about 1/64 millionth of a second, compared to seconds for traders in the 1980s.
These trades are typically very short term in nature and may buy or sell the same security multiple times within a nanosecond. AI algorithms analyze massive amounts of data in real time to identify patterns and trends that are not immediately apparent to human traders. This allows traders to make better decisions and execute trades at a faster pace than they would be able to do it manually.
Another important application of AI in HFT is natural language processing. This includes analyzing and interpreting human language data such as news articles and social media posts. By analyzing this data, traders can gain valuable insight into market sentiment and adjust their trading strategies accordingly.
Advantages of AI trading
These AI-based high-frequency traders work in a very different way than humans do.
The human brain is slow, imprecise, and forgetful. It lacks the fast and precise floating-point math required to analyze large amounts of data to identify trading signals. Computers are millions of times faster, with essentially solid memory, impeccable attention, and infinite capabilities that allow them to analyze large amounts of data in milliseconds.
So, like most technologies, HFT offers some advantages to the stock market.
These traders usually buy and sell assets at prices very close to market prices. This means that we do not charge high fees to our investors. This ensures that there are always buyers and sellers in the market, thus stabilizing prices and reducing the possibility of sudden price fluctuations.
High-frequency trading also helps mitigate the effects of market inefficiencies by quickly identifying and exploiting market mispricing. For example, HFT algorithms can detect whether a particular stock is undervalued or overvalued and take advantage of these discrepancies to execute trades. By doing so, this type of trading helps correct market inefficiencies and more accurately determine the price of an asset.
Disadvantage
But speed and efficiency can also do harm.
The HFT algorithm reacts very quickly to news events and other market signals, which can cause sudden spikes and falls in asset prices.
Additionally, HFT financial firms can use their speed and technology to gain an unfair advantage over other traders and further distort market signals. The volatility created by these highly sophisticated AI-powered trading beasts caused the so-called Flash Crash in May 2010. Stocks plummeted then recovered within minutes, wiping out about $1 trillion in market value and then recovered.
Since then, volatile markets have become the new normal. In a 2016 study, two co-authors and I found that volatility (a measure of how quickly and unpredictably prices go up and down) increased significantly after the introduction of HFT.
The speed and efficiency with which high-frequency traders analyze data means that even small changes in market conditions can trigger large numbers of trades, leading to sudden price movements and increased volatility.
Additionally, a study I published in 2021 along with several other colleagues showed that most high-frequency traders use similar algorithms, increasing the risk of market failure. This is because as the number of these traders increases in the market, the similarity of these algorithms may lead to similar trading decisions.
This means that all high-frequency traders are likely to trade on the same side of the market if the algorithm emits similar trading signals. That is, they may all be willing to sell in the case of negative news and buy in the case of positive news. Without someone standing on the other side of the trade, the market can fail.
Enter ChatGPT
This gives us a new world of trading algorithms and similar programs powered by ChatGPT. They pick up the issue of too many traders on the same side of the trade and could make things even worse.
In general, humans tend to make various decisions at their own discretion. However, diversity of opinion may be limited if everyone derives their decisions from similar artificial intelligence.
Consider the extreme non-financial situation where everyone relies on ChatGPT to determine the best computer to buy. Consumers are already herding behavior where they are more likely to buy the same products and models. For example, reviews from places like Yelp and Amazon motivate consumers to choose among some of the top options.
Because decisions made by chatbots powered by generative AI are based on past training data, similarities exist in the decisions suggested by chatbots. It is very likely that ChatGPT will suggest the same brand and model to everyone. This could take grazing to a whole new level, potentially leading to shortages and severe price hikes for certain products and services.
This problem is exacerbated when the decision-making AI is informed by biased and inaccurate information. AI algorithms can reinforce existing biases if the system is trained on biased, outdated, or limited datasets. ChatGPT and similar tools have also been criticized for being factually incorrect.
Moreover, market crashes are relatively rare, so there isn’t much data on them. Because generative AI relies on data training to learn, lack of knowledge about generative AI makes it more likely to occur.
At least for now, most banks don’t seem to allow their employees to use ChatGPT or similar tools. Citigroup, Bank of America, Goldman Sachs and several others have already banned their use on trading room floors, citing privacy concerns.
But I strongly believe that banks will eventually adopt generative AI once their concerns about generative AI are resolved. The potential benefits are too great to ignore and the risk of being left behind by your competitors.
However, the risks to financial markets, the global economy, and to everyone are also great, so I urge you to act with caution.
This article originally appeared in The Conversation. Please read the original article.
