JPMorgan has developed a ChatGPT-based language model designed to analyze Federal Reserve statements and speeches, aiming to identify potential trading signals.
The model has reviewed 25 years’ worth of Federal Reserve speeches, ranking them according to a “Hawk-Dove” score, as reported by Bloomberg. JPMorgan’s economists have then compared this index against asset performance, suggesting that the AI could assist in predicting policy changes and could be beneficial for trading strategies.
In a recent note, the bank indicated that “preliminary applications are encouraging,” and the model has already been adapted to include the European Central Bank and the Bank of England, with plans to extend its functionality to more central banks in the future.
While ChatGPT holds significant promise in financial services, firms like JPMorgan have been cautious in its implementation. Earlier this year, it was reported that the bank is limiting employees’ use of the AI chatbot due to concerns about bias, accuracy, and an evolving regulatory environment.
In his annual letter to shareholders, JPMorgan Chase CEO Jamie Dimon disclosed that the bank has over 300 active AI use cases, describing the technology as “extraordinary and groundbreaking.”