Two new research papers published this month demonstrate the potential of OpenAI's ChatGPT in the world of finance.
The AI chatbot was put to the test in deciphering Federal Reserve statements and determining whether headlines were good or bad for a stock, and it aced both tests, according to a Bloomberg report.
The first study, titled "Can ChatGPT Decipher Fedspeak?", was conducted by researchers from the Federal Reserve who found that ChatGPT outperformed commonly used models from Google and classifications based on dictionaries. It also explained its classifications of Fed policy statements in a way that resembled the central bank's own analyst.
The second study, "Can ChatGPT Forecast Stock Price Movements?", prompted ChatGPT to interpret corporate news headlines and found that the answers given by the chatbot showed a statistical link to the stock's subsequent moves, demonstrating its ability to correctly parse the implications of news.
These findings suggest a potentially major step forward in the use of technology to turn reams of text into trading signals. While the use of language models in finance is nothing new, the advances demonstrated by ChatGPT show a new level of parsing nuance and context.
The research also suggests that ChatGPT can pull off similar tasks without being specifically trained, indicating the potential of "zero-shot learning" to exceed prior technologies. The technology could open up new worlds of information and make it more accessible to a broader community of finance professionals.