The new risks ChatGPT poses to cyber security - 2024

Divine Translation Bureau is a compilation team under 36 Krypton, focusing on technology, business, workplace, life, and other fields, focusing on introducing new technologies, new ideas, and new trends from abroad.

Editor's note: In the financial world, with the development of technology and technology, transactions have become more complex and high-frequency. History has proved that the more advanced the technology, the greater the market volatility. In this process, there are beneficiaries and there are victims. This article is from a compilation, I hope to inspire you. 

• AI-powered tools, such as ChatGPT, have the potential to revolutionize the efficiency, effectiveness, and speed of human work.
• This is true in financial markets, but it is also true in healthcare, manufacturing, and just about every other aspect of our lives.

I have studied financial markets and algorithmic trading for 14 years. While artificial intelligence offers many benefits, the growing ubiquity of these technologies in financial markets also brings potential dangers. Looking at Wall Street's past attempts to speed up trading by embracing computers and artificial intelligence, we can spot some important lessons about using these technologies for decision-making.

Programmatic trading spawned "Black Monday"

In the early 1980s, spurred by technological advances and financial innovations such as derivatives, institutional investors began using computer programs to execute trades based on pre-set rules and algorithms. This helps investors complete large transactions quickly and efficiently.

At the time, these algorithms were relatively simple and were mainly used for so-called index arbitrage, which is to profit from the difference between the price of "a stock index such as the S&P 500" and the "stocks that make up the index".

As technology advances and more data becomes available, this programmatic trading becomes more sophisticated and algorithms begin to analyze complex market data and execute trades based on various factors. The number of these programmatic traders continues to grow on the largely unregulated trading highway, with more than $1 trillion worth of assets changing hands every day, leading to a sharp increase in market volatility.

Ultimately, this led to the massive stock market crash of 1987, known as Black Monday. The Dow Jones Industrial Average suffered its worst drop ever, and the pain spread across the globe.

In response, regulators have implemented a series of measures to limit the use of programmatic trading, including circuit breakers and other restrictions that suspend trading during major market fluctuations. But despite these steps, programmatic trading has continued to gain popularity in the years following the crash.

High-Frequency Trading (HFT)

Fifteen years later, in 2002, the New York Stock Exchange launched a fully automated trading system. As a result, programmatic traders gave way to more sophisticated automated trading and a more advanced technique: high-frequency trading.

High-frequency trading uses computer programs to analyze market data and execute trades at extremely high speeds. Unlike program traders, who take advantage of arbitrage opportunities by buying and selling baskets of securities over long periods of time, high-frequency traders use powerful computers and high-speed networks to analyze market data and execute trades at lightning speed. High-frequency traders can make trades in about 64 millionths of a second, compared with the seconds it took traders in the 1980s.

These trades are typically very short-term and may involve buying and selling the same security multiple times within nanoseconds. AI algorithms are able to analyze large amounts of data in real-time and identify patterns and trends that human traders cannot see instantly. This helps traders make better decisions and execute trades faster than manually.

Another important application of artificial intelligence in high-frequency trading is natural language processing, which involves analyzing and interpreting data in human language, such as news articles and social media posts. By analyzing this data, traders can gain insights into market sentiment and adjust their trading strategies accordingly.

Benefits of AI Trading

These artificial intelligence-based high-frequency transactions operate very differently from human transactions.

The human brain is sluggish, inaccurate, forgetful, and incapable of fast, high-precision floating-point arithmetic, a skill required to analyze large amounts of data to identify trading signals. But computers are millions of times faster than the human brain, with impeccable memory, perfect focus, and an unlimited ability to analyze vast amounts of data in milliseconds.

So, like most technologies, high-frequency trading brings several benefits to the stock market.

High traders typically buy and sell assets very close to market prices, which helps ensure that there are always buyers and sellers in the market, which in turn helps stabilize prices and reduce the likelihood of price spikes. 

 Fast trading can also help reduce the impact of market inefficiencies by quickly identifying and exploiting market mispricing. For example, high-frequency trading algorithms can detect when a particular stock is undervalued or overvalued and execute trades to take advantage of these differences. Such transactions can help correct market inefficiencies and ensure more accurate pricing of assets. 

Disadvantages of AI trading 

But speed and efficiency can also hurt markets.  High-frequency trading algorithms can react very quickly to news and other market signals, causing sharp jumps or falls in assets. In addition,  financial firms that trade frequently can use their speed and technology to gain an advantage over other traders, which in turn further distorts market signals. The volatility caused by these highly sophisticated AI-driven trades led to the so-called flash crash in May 2010, when stocks fell and then recovered within minutes, wiping about $1 trillion off the market capitalization before Jai's quick recovery. 

In a 2016 study, two authors and I found that volatility (a measure of the speed and unpredictability of price rises and falls) increased significantly after the introduction of high-frequency trading. 

 

The speed and efficiency with which high-frequency traders analyze data mean that even small changes in market conditions can trigger huge trading volumes, causing sudden price changes. 

This is because the similarity of their algorithms leads to similar trading decisions when the number of traders in the market increases.  This means that all high-frequency traders are likely to trade on the same side of the market if their algorithms send similar trading signals. In other words, they are all probably trying to sell negative news and buy positive news. If no one is on the other side of the deal, the market fails.

Enter the ChatGPT Era  

Artificial intelligence has brought us into a new world of trading algorithms produced by ChatGPT and similar programs. And these techniques can exacerbate the "too many traders on the same side" problem. In general, people tend to make different decisions when they let nature take its course. But having all decisions about the same AI can limit the diversity of opinion.  Consider an extreme, non-financial situation where everyone depends on ChatGPT to decide which computer is the best to buy. Nowadays, consumers are already very prone to herd behavior and tend to buy the same product and model. 

For example, reviews on sites like Yelp, Amazon, etc. make consumers choose from several main options.  Since the decisions made by a  generative AI chat are based on previous training information, there are similarities in the decisions the chat offers. It is possible that ChatGPT recommends the same make and model for everyone. This could raise the "herd effect" to an even higher level and lead to shortages of certain products and services and serious price increases. 

This becomes even more problematic when AI makes decisions based on biased and incorrect data. When systems are trained to use misleading, outdated, or limited data sets, AI algorithms reinforce existing biases. ChatGPT and similar tools have been widely criticized for factual errors. Also, because market crashes are relatively rare, there isn't much information about them. Because generative AIs rely on data training to learn, their ignorance can make crashes more likely. 

Employees to use ChatGPT

Most banks, at least for now, do not allow employees to use ChatGPT and similar tools. Citigroup, Bank of America, Goldman Sachs, and several other banks have banned the tools from their trading floors, citing privacy concerns.  But I strongly believe that if banks address their concerns about generative AI, they will eventually adopt generative AI. Because the potential gain is too great to lose, and you risk being left behind by your competitors.  But there are also significant risks for financial markets, for the global economy, for everyone, so I want them to proceed with caution.

How might artificial intelligence affect the risk of nuclear war?