Delphia was fined $225K by the SEC for AI washing [1]. This incident is just a glimpse of a much larger iceberg looming in the financial waters, an iceberg compounded by rising interest rates. When the tide is high, and the cost of money is zero, distinguishing between genuine innovation and slick pitch deck becomes increasingly challenging. Bias is a strange beast. When you're biased towards something, objectivity flies out the window. This explains why women, blacks, and immigrants remain underfunded at all times.
Now that the tide has gone out, the regulators are coming. They understand that the march of machines is inevitable, but if they don't start segregating the wheat from the chaff now, they're staring down a colossal problem.
The current situation reminds me of Harry Markopolos [2], who raised thirty red flags to expose that the world's biggest hedge fund was a scam. Yet, even his persistence needed the catastrophic crash of 2008 to bring the issue to light. Any stock market crash is going to make things tricker for the washers.
I'm no Markopolos, but I've sounded the alarm, not once [3], not twice [4], but multiple times [5][6][7][8][9]. Alpha generation through crowdsourcing is a myth. Yet, the fear of missing out (FOMO) drives investment decisions more often than rational thought.
Moreover, who has time to ponder over Charles Mackay's [10] insights when there's a promise that machines can extract wisdom from crowds—crowds that, by their very nature, are predisposed to madness? The wisdom of crowds can fail in not just one way, but many. And for similar reasons, no matter how hard you try, AI can't beat the markets.
Here's the crux of my argument:
AI is a computational tool, not a statistical innovation.
It has an informational constraint. If the information fed into it is garbage, expecting the outcome to be gold is folly.
AI is utterly clueless about complexity. Markets embody complexity. Statistics fall short in explaining complexity, and yet, a generation of stock market participants has been schooled into believing that the market has nothing to do with physics by economists and behavioral finance scholars [11][12].
Some articles I wrote warning about AI and explaining why stock markets are governed by physics [13]. Beating markets by LLM is a fool’s game [14]. Need no AI to beat the market [15]. My Q&A with chatGPT [16].
So here are your reasons to tread carefully with AI. Whatever you do, if you don't think fundamentally different, you are going back to square one.
Innovation doesn't need billions of dollars or excessive computation; it requires common sense and thinking outside the box, which, generally, lies outside the human fat tail.
All I can wish you then is good luck, in case you end up being AI washed. Oh! And by the way, the other industry that relies on wisdom of crowds, it’s called passive investing. Good luck with that too!
[2] https://www.linkedin.com/pulse/worlds-largest-hedge-fund-fraud-mukul-pal-caia/
[3] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3070978
[5] https://www.linkedin.com/pulse/paradox-mediocrity-mukul-pal/
[6] https://www.linkedin.com/pulse/human-ai-mukul-pal-caia/?trackingId=U%2FuSQ92kRam8uFu2zp9mbw%3D%3D
[8] https://mukulpal.com/blog/2017/3/31/architecture-of-data?rq=wisdom%20of%20crowds
[9] https://mukulpal.com/blog/2023/8/2/noise?rq=wisdom%20of%20crowds
[10] https://www.linkedin.com/pulse/extraordinary-delusions-mukul-pal-caia/?trk=articles_directory
[11] https://arxiv.org/pdf/1003.2688.pdf