In a brochure published this week, IRIS researcher Caroline Joly shows that artificial intelligence (AI) is responsible for more than two-thirds of decision-making worldwide regarding stock buys and sells. It’s called algorithmic trading. We’ll probably hear “Who cares so long as it works?” The brochure nonetheless demonstrates that such a significant transformation has important consequences for us.
We imagine the stock exchange packed with traders shouting out their buy and sell orders, but the reality has changed a lot, explains Joly. At the moment, firms operating on the financial markets are all about who has the best algorithm, the most powerful computer, or the shortest distance separating its servers from those of the biggest stock exchanges. These robots can take advantage of minuscule price fluctuations visible for microseconds only. This is called high-frequency trading. In such situations, no one can compete with machines which exchange more than a thousand securities in a wink of an eye.
High-frequency AI-generated transactions are not free from economic risk. Within just a few minutes, perhaps in a few seconds, they can lead to huge monetary losses, bankruptcies and even job losses.
For example, there was a “flash crash” on May 6, 2010. In the span of a few minutes, the robots took away close to a trillion dollars from the markets, putting businesses on the brink of bankruptcy. More recently, on April 23 last, a robot analyzed a fake tweet stating that the White House was under attack and that President Obama had been injured. As a result, all robots tried to sell their securities and in little more than three minutes Wall Street had lost 136 billion dollars.
The brochure also explains why financial markets need robots so bad nowadays. Financial products have simply become too complex for human beings. We not only need to understand them, but also be able to know what are the optimal circumstances in which to use them. Our capabilities are just no match for AI. It all stems from the financial deregulation of the 80s, during which period transactions multiplied exponentially. Ever more complex products were created to manage risk and to stabilize the financial markets. Today, “financial innovation” has largely exceeded its creators. For the better? That’s no sure bet.
This article was written by Guillaume Hébert, a researcher with IRIS—a Montreal-based progressive think tank.