AI in Hedge funds: Bridgewater trusts ChatGPT to pass investment associate test

In a significant development that highlights the increasing role of artificial intelligence (AI) in finance, Bridgewater Associates, one of the world's largest hedge fund by assets under management (AUM), recently announced that OpenAI's ChatGPT successfully passed its investment associate test. This achievement is especially notable considering Bridgewater Associates' status as a trailblazer in the hedge fund industry, boasting $150 billion AUM and offering a range of innovative strategies to institutional clients.

AI in the investment process

The revelation implies that AI, specifically ChatGPT, possesses the ability to perform financial analysis and make investment decisions at a level comparable to a human investment associate. This capability can potentially transform the industry by increasing efficiency, reducing costs, and even improving the accuracy of financial predictions.

The test taken by ChatGPT likely examined the AI's ability to perform the critical tasks of an investment associate. These tasks can include financial modelling, risk analysis, asset valuation, and more. Considering the high average salary for a hedge fund the use of AI in these roles could result in substantial cost savings for hedge funds worldwide.

What is Bridgewater’s plan

Bridgewater's statement about 'having millions of them at once' suggests that ChatGPT's scalability is a significant advantage. Unlike human analysts who can only work a certain number of hours a day and manage a finite number of tasks, AI can be deployed at scale, providing around-the-clock service and handling an unlimited number of tasks simultaneously.

This scalability can drastically increase the capacity of hedge funds to analyze a greater quantity of data and make more informed investment decisions. It also allows funds to explore new strategies that would be too resource-intensive for human teams.

While the world's largest hedge sees great potential in AI technologies like ChatGPT, it's important to consider different perspectives from within the firm. Greg Jensen, co-chief information officer at Bridgewater Associates, recently shared his views on AI's role in the stock market.

The AI Bull with a warning

Jensen, who is an AI enthusiast and early investor in OpenAI, the creator of ChatGPT, warned against the sole use of large language models (LLMs) like ChatGPT for stock trading. Speaking on the 'Odd Lots' podcast from Bloomberg, Jensen said that choosing LLMs for stock picking was "hopeless.

"If somebody's going to use large language models to pick stocks, that's hopeless, that is a hopeless path". He added that such technology would be "awful" at identifying major shifts in the macroeconomic environment.

This viewpoint doesn't mean he entirely dismisses AI's place in wealth management. In fact, Bridgewater Associates, under the leadership of billionaire investor Ray Dalio, has been utilising machine learning for a while. Jensen himself is a proponent of harnessing AI's strengths by integrating various models into their strategies.

Limitations of ChatGPT's investment advice

Despite its impressive capabilities, ChatGPT has limitations when it comes to providing investment advice. For instance, as of its knowledge cutoff in September 2021, ChatGPT lacks access to current market trends or events that have occurred post-2021. If asked about recent events, it typically responds with a note about this cutoff, which is a clear limitation for real-time investment advice.

Also, ChatGPT doesn't have access to certain industry-specific information or up-to-date data relevant to specific investment opportunities. For example, it cannot compare the most recent financial statements and earnings reports, measure current trends in consumer spending, or evaluate government policies—elements that significantly impact a company's profitability.

Although trained on an extensive amount of data, ChatGPT has been known to provide incorrect responses, highlighting the need for additional user research to ensure accuracy.