The Blubber

The finance industry is witnessing the emergence of academic research that applies ChatGPT, an artificial intelligence (AI) chatbot, to finance-related tasks. Early results indicate that the technology lives up to its hype.

Two recent studies have demonstrated ChatGPT's capabilities in finance. The first study assessed ChatGPT's ability to interpret whether Federal Reserve statements were hawkish (suggesting tighter monetary policy) or dovish (suggesting looser monetary policy). The second study evaluated ChatGPT's capacity to determine the impact of news headlines on stock prices. In both cases, ChatGPT performed impressively, indicating a significant advancement in the use of technology to convert textual data, such as news articles, tweets, and speeches, into actionable trading signals.

While quantitative analysts (quants) on Wall Street have long employed language models, including natural language processing (NLP), to inform their strategies, ChatGPT's performance suggests that the technology developed by OpenAI has reached a new level of sophistication in understanding context and nuance.

Slavi Marinov, head of machine learning at Man AHL, which has used NLP to analyze texts like earnings transcripts and Reddit posts, remarked, "It's one of the rare cases where the hype is real."

The first study, titled "Can ChatGPT Decipher Fedspeak?," was conducted by Anne Lundgaard Hansen and Sophia Kazinnik from the Federal Reserve. They found that ChatGPT outperformed Google's BERT model and dictionary-based classifications in interpreting the central bank's statements. Notably, ChatGPT was able to provide explanations for its classifications that closely resembled those of a human analyst.

In the second study, titled "Can ChatGPT Forecast Stock Price Movements? Return Predictability and Large Language Models," researchers Alejandro Lopez-Lira and Yuehua Tang from the University of Florida used ChatGPT to analyze corporate news headlines. The study found a statistical correlation between ChatGPT's interpretations and subsequent stock price movements, indicating the chatbot's ability to accurately understand the implications of news.

The use of NLP to analyze stock popularity on Twitter or to process company headlines is now commonplace among advanced quants. However, ChatGPT's advancements are expected to unlock new sources of information and make the technology more accessible to a wider range of finance professionals.

Marinov believes that while it is unsurprising that machines can now read almost as well as humans, ChatGPT has the potential to accelerate the entire process. Man AHL previously relied on manual labeling of sentences as positive or negative for an asset to train its models. The firm later gamified the process to involve all employees. The new studies suggest that ChatGPT can perform similar tasks without specific training, a capability known as zero-shot learning. This capability already surpasses previous technologies and can be further improved with fine-tuning.

You can use unusualwhales ai tools at https://unusualwhales.com/ai