CONSOLIDATION AND CASCADING COLLAPSE: DANGERS OF AI IN FINANCE

0
273

by Joseph P. Farrell, Giza Death Star:

This was another of those weeks when there was a surfeit of stories to blog about, and I wish I could blog about them all. There is, of course, a veritable tidal wave of stories and speculations about the sad mid-air collision at Reagan National Airport, including an excellent article about complex systems and the fragility of the USA, especially with the introduction of “Diversity, Inclusion, and Equity” policies. Many of these I have tried to put in this week’s “Honourable Mentions.”  With all this news, there has also been news about artificial intelligence, and some very intriguing and suggestive moves regarding gold this week, moves that almost compel speculation.  So, lest these stories drop off the radar or we lose their transponder signals, I’ve decided to devote this week’s blogs to that theme.

TRUTH LIVES on at https://sgtreport.tv/

With that in mind, V.T. spotted this article over at our friends at Zero Hedge, and sent it along (thank you!). The dangers it points out a inherent in the rush to turn equities and commodities trading into more complex systems. For some time, I’ve been warning and commenting that the race to more and more algorithmic trading and “dark pools” would, eventually, lead to a market driven by forces out of touch with human risk and investment assessment. In this regard, in blogs over the past few years, I’ve pointed to the various “flash crashes”  in some stocks that algorithmic trading has led to, and warned that at some point, this phenomenon would not be restricted to one or two stocks or commodities, but become potentially a market-wide phenomenon.

Now, it seems, Jim Rickards is raising a similar alarm, and his argument is a rather nuanced and subtle one:

The Hidden Dangers Of AI In Finance

Rickards’ argument is that the increasing use of algorithmic trading and the adoption of Artificial Intelligence into such algorithmic trading could produce a cascade effect; here’s Rickards’ actual words:

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.

To this scenario, Zero Hedge adds its own details:

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?

This means that the release of a new cutting-edge model could actually change the market’s direction. It’s newer, it’s smarter, and it thinks stocks are overvalued by 40%! Sell!

With the mention of China’s new “DeepSeek” Artificial intelligence, I suggest there is yet a third, and very obvious, danger.  Our friend and colleague Catherine Austin Fitts has said – repeatedly and in many interviews with many people over the years – that there is no cyber system that is secure. With the advent of DeepSeek, we’re in truly unchartered territory, for already stories have appeared of people asking DeepSeek for information on how many people Mao Tse Tung killed. The result was an initial return that mentioned Mao’s genocidal tendencies. But within mere seconds, the AI had erased its initial response, and offered a bland view of Mao as a sort of Chinese Mr. Rogers. The person making the inquiry, however, kept a screen capture of the initial response, and then confronted the AI with it, asking why it had “cleansed” it and removed it. The “AI” had no recollection of having done so.

All this to my mind poses the real danger with algorithmic trading coupled with the new generations of AI: the financial markets are exposed to infiltration by hostile AIs, which could conceivably take down whole trading networks, and then  falsify the trading record itself or, worse, simply “forget” what trade had been executed. Mao? He didn’t execute anyone. He was a sweet roly poly little dumpling of a man with dimples. So sorry you lost fifteen trillion dollars of assets, but we have no memory nor records of such a loss.

But here’s the clincher. In the Zero Hedge article, there is at the very end an intriguing admission, one that I, as an analogue advocate have been harping on for years: nothing is not a substitute for physical media and physical records, and by “Nothing” I mean precisely any form of blips on a computer screen passing for the “foundational money” or “currency of transaction.”  At the end of the Zero Hedge article, there is this welcome admission:

Read More @ GizaDeathStar.com