> A whopping, 50,000 employees have access to JPM’s proprietary AI chatbot. > > An internal memo seen by the FT, says LLM Suite can help employees in the asset & wealth divisions with writing, idea generation and summarizing documents. > > “Think of LLM Suite as a research analyst that can offer information, solutions and advice on a topic”. > > JPM will spend around $17 billion on technology this year and already have 2,000 AI/ML experts and data scientists working for them. > > In a recent interview with BBG, CEO Jamie Dimon said: > > “Every single process: errors, trading, hedging, research, every app, every database you’re going to be pulling AI… > > AI’s doing all the equity hedging for us, for the most part”. > > Wall Street rival, Morgan Stanley also launched its own AI assistant for wealth advisors “Debrief” this year. > > One thing is for certain — the top firms on Wall Street are investing heavily in AI and are already rolling out solutions.
The thing is that LLMs require data generated by humans to learn and the [more they rely on synthetic results, the more they contaminate their own training data](https://www.techtarget.com/whatis/feature/Pros-and-cons-of-AI-generated-content). For example, with image generation systems, the more pictures of humans with six fingers you accept (because the rest of the picture looks nice enough), the more the underlying neural network will think humans have that many fingers… until the results are garbage.
If these firms rely on LLMs too much, soon they are going to run out of new, reliable data to keep the systems updated. And much like European Royalty, they will suffer of “consanguinity” issues until they become unfit
These firms want to announce the use of AI to scare their employees into accepting worse working conditions, but they are aware of these inherent limitations of the technology and that they need experts to validate the results. Probably less than what they have now… but, at the same time, they should be scared of what those experts might do if they fire them, because, as always, technology is a double-edged sword. The more accessible these AI systems become, the easier it is to create your own company and compete against established firms.
2 comments
> A whopping, 50,000 employees have access to JPM’s proprietary AI chatbot.
>
> An internal memo seen by the FT, says LLM Suite can help employees in the asset & wealth divisions with writing, idea generation and summarizing documents.
>
> “Think of LLM Suite as a research analyst that can offer information, solutions and advice on a topic”.
>
> JPM will spend around $17 billion on technology this year and already have 2,000 AI/ML experts and data scientists working for them.
>
> In a recent interview with BBG, CEO Jamie Dimon said:
>
> “Every single process: errors, trading, hedging, research, every app, every database you’re going to be pulling AI…
>
> AI’s doing all the equity hedging for us, for the most part”.
>
> Wall Street rival, Morgan Stanley also launched its own AI assistant for wealth advisors “Debrief” this year.
>
> One thing is for certain — the top firms on Wall Street are investing heavily in AI and are already rolling out solutions.
The thing is that LLMs require data generated by humans to learn and the [more they rely on synthetic results, the more they contaminate their own training data](https://www.techtarget.com/whatis/feature/Pros-and-cons-of-AI-generated-content). For example, with image generation systems, the more pictures of humans with six fingers you accept (because the rest of the picture looks nice enough), the more the underlying neural network will think humans have that many fingers… until the results are garbage.
If these firms rely on LLMs too much, soon they are going to run out of new, reliable data to keep the systems updated. And much like European Royalty, they will suffer of “consanguinity” issues until they become unfit
These firms want to announce the use of AI to scare their employees into accepting worse working conditions, but they are aware of these inherent limitations of the technology and that they need experts to validate the results. Probably less than what they have now… but, at the same time, they should be scared of what those experts might do if they fire them, because, as always, technology is a double-edged sword. The more accessible these AI systems become, the easier it is to create your own company and compete against established firms.