Some disturbing facts about AI and banking that can have profound consequences

Back in November I wrote an essay entitled Wasting a trillion dollars is terrible. Even then, it was clear that generative AI would probably lose a lot of money. While there is some value in this, the costs are high and the revenues have been modest relative to these costs. The big problem that has existed since the early days of LLMs is that reliability is not great and therefore the return on investment has been low for most large enterprise customers (as several studies have now shown).

What I couldn’t imagine at the time was how dire the consequences could be if everything fell apart. I pointed out the vulnerability of the banking sector as a key unknown (bold added):

But it’s not just a matter of a trillion dollars or more going down the drain, there could also be significant collateral damage to the rest of society, both economic and otherwise (e.g. in terms of how LLMs have undermined higher education). As…