AI seems to be all anyone can talk about these days, and the “AI arms race” is off and running, without much apparent concern as to the consequences.
At least one SEC head fears that one of those consequences could be a huge market crash – which obviously wouldn’t be good for anybody.
When SEC chair Gary Gensler was still a MIT professor he wrote a paper arguing that a bunch of too-similar deep learning programs could undermine our economic structures and cause a real crisis.
He and coauthor Lily Bailey thought a broad adoption of AI would tip over the fragile uniformity and interconnectedness of our economic systems and leave our financial systems vulnerable.
They also worried that our human-speed financial technology and analytics would not be able to keep up.
Their main issue is that they feared training AI models on the same data would lead to “highly correlated predictions that proceed in lockstep.”
Basically, if all of the advanced AI models came to the same conclusions, bad things would happen.
The warnings also go into machine bias and the reality that AI systems are inevitably full of human social prejudices.
“This technology will be the center of future crises, future financial crises. It has to do with this powerful set of economics around scale and networks.”
Meaningful regulatory action hasn’t been taken at this time, but with this man at the helm of at least one company, we can hope that it will be.
Or it sounds as if we might all pay the price.