More than 70% of all stock market activity is now being traded by thinking machines, and many fear this could spell a cataclysmic event if the system begins to fall apart.
The stock market has certainly enjoyed a year of record highs, and a performance that seemed to tick ever-upward. However, while most retail investors do not specifically select their own stocks, they probably never considered that a robot was doing it for them, let alone that there’s a smarter one on the other side of virtually every trade.
The former head of Nasdaq estimates that less than 30% of all stock trades are executed by actual humans. This means that more than 70% of everything else can be attributed to some kind of machine, robot or computer-generated algorithm.
High Frequency Trading (HFT) tends to be the most pervasive method, often operating under the radar to nefariously manipulate the stock market. HFT has also been known to exacerbate systematic market risk, having played a highly suspect role in the trillion dollar 2010 Flash Crash. Essentially, HFT is an automatic trading program that is governed by algorithms to study data and execute trades in massive quantities at incredibly high speeds. Unfortunately, the discreet nature of their methods and trading strategies make its true market size almost impossible to know.
But while HFT certainly poses a very real threat to the financial system, some believe there looms an even greater danger.
That danger lies in the market’s presence of Artificial Intelligence.
Leading edge AI models are now being designed with a neural architecture inspired by the fields of genetic evolution and probabilistic logic. For example, Sentient Technologies, a San Francisco-based hedge fund, has developed a system of AI that uses evolutionary computation. Derived from genetics, it uses thousands of machines running simultaneously around the world, algorithmically creating trillions of virtual traders that it calls “genes.” These genes are tested with hypothetical sums of money to trade in simulated situations created from historical data. The genes that are unsuccessful die off, while those that are profitable are spliced together with others to create the next generation.
And deep within Sentient’s offices, eerily adorned with “Terminator” movie posters, sit two men in a windowless trading room who quietly monitor the system at all times. There’s even an emergency red button “if all hell breaks loose.”
Sentient’s co-founder, Babak Hodjat, is a computer scientist who helped lay the groundwork for Apple’s Siri. He also is possessed by the unbalanced notion that logic transcends intuition, and has created a sentient machine operating under the same principles. ”For me, it’s scarier to be relying on those human-based intuitions and justifications than relying on purely what the data and statistics are telling you,” says Hodjat.
But isn’t intuition what gives us our very humanity?
It is often said that in the end, the world of investing boils down to two basic emotions: greed and fear. A machine that thinks purely in data may be appealing for financial gain, but what happens as it begins to learn and evolve at a geometric rate?
Has no one considered what might happen when a highly evolved binary thinking machine, with a complete lack of human emotion, is suddenly at the helm of a globally interdependent financial economy?
In other words, what happens when that emergency red button no longer exists?
Dangers to the Financial Markets
The fact is, this terrifying possibility of disaster stems from the public’s false assumption that AI merely follows the same rules of programming that a human would follow if presented with the same data.
“Most people think of artificial intelligence…as simply executing logical rules programmed into them by humans,” says Doug Kass, President of Seabreeze Partners Management. Kass is a hedge fund manager with serious concerns about the use of these technologies, and has begun sounding the alarm. “The general belief is that…AIs are just ‘faster humans able to do a lot more calculations in a meaningful time frame.’ That may NOT be a correct characterization of some of the more powerful AIs that may be working in the markets.”
Which leads us to another problem: no one knows just how much financial wherewithal is in the hands of AI, because no regulation requires it to be disclosed.
According to Kass, this is a “very, very big gap in regulation.”
In other words, the full impact of AI on the financial system may not be known until its far too late. This is also why understanding the key difference between High Frequency Trading and Artificial Intelligence is so important.
AI is not merely an innocuous software program that behaves as its human programmer would, nor does it facilitate predictable behaviors based upon a given set of data and conditions. Genuine artificial intelligence will learn, think and ultimately make decisions for itself. Herein lies the danger in unleashing Pandora’s Box upon the worldwide financial system.
Kass remains insistent on this final point: ”The creators have no knowledge of what their creations are thinking…the machines are inscrutable and, most terrifyingly important, UNPREDICTABLE.”