Jim Rickards recently published a compelling article on AI risk for Insider Intel subscribers.
In it, Jim discusses a different way in which AI could crash markets. One that is totally separate from the DeepSeek, China, and NVIDIA angle we’ve been covering for the past week.
Today we’re going to review his key points and explore them in detail.
Here’s Jim:
“The ultimate danger arises when a large cohort of asset managers controlling trillions of dollars of assets all employ the same or similar AI algorithms in a risk management role. An individual robot working for a particular asset manager tells the manager to sell stocks in a crashing market. In some cases, the robot may be authorized to initiate a sale without further human intervention.
Taken separately, that may be the best course of action for a single manager. In the aggregate, a selling cascade with no offsetting buy orders from active managers, specialists or speculators takes stock prices straight down. Amplification through feedback loops makes matters worse.
Individual AI systems have various trigger points for selling. Not all will be triggered at once, yet all will be triggered eventually as selling begets more selling, which triggers more automated systems that add to the selling pressure, and so on. There are no contrarians among the robots. Building sentiment into systems is still at a primitive stage.”
This is a good example of why I read Jim’s work. He always approaches issues from a unique and thoughtful angle.
This risk is clearly real. We are now at the point where trading firms are integrating LLMs (AI models) into their proprietary algorithms.
What happens if a majority of trading firms are using the same AI software to drive their trading? For example, it’s likely that many money managers have integrated OpenAI’s ChatGPT models into their algos.
Now that DeepSeek R1 is the new shiny object, maybe a significant portion of firms are switching to that model.
Perhaps DeepSeek approaches trading in a completely different way. What happens if ChatGPT interprets data bullishly, but DeepSeek sees the same information as bearish?
Hellbound and Down 