Can ChatGPT Predict Stock Market Moves? A New Study Says Yes

Hundreds of millions of white-collar jobs are set to be displaced by artificial intelligence in the near future—too near, really. One of them could well be investment analysts. According to research conducted by finance professors from the University of Florida, ChatGPT already possesses the capacity to forecast its own stock market movement predictions.

The professors had fed the chatbot more than 50,000 news headlines about various conglomerates from October 2021 to the present. It would go on to evaluate the news' correlation with the companies' stock prices. In the end, it came up with what's called a ChatGPT score by utilizing sentiment analysis. The score was assessed to see if it could predict the groups' stock market performance the next day.

Of course, it did. The study concluded that there was a relationship between the scores and the analyzed companies' performances the following day. Companies that had higher scores were more than likely to have better returns while those with lower scores had worse ones. But the real kicker here, however, is the fact that ChatGPT was said to have outperformed traditional sentiment analysis methods that followed the same model of predictive stock movement data.

ALSO READ:

We Interviewed ChatGPT, the AI That's Making Google Very, Very Nervous

Ryan Reynolds Asked ChatGPT to Write an Ad Using His Voice. The Result Was 'Mildly Terrifying'

This proves that ChatGPT and investment decision-making make sense and can get us more accurate predictions. It can also mean improved performances across the board in terms of quantitative trading strategies. The study itself is a promising lead for investors who want to get a better read on stock market movements in due time.

There are, however, apprehensions about such. Although if we were to ask Bloomberg, it might be a different story. Its new GPT-based language model BloombergGPT was trained on a particular dataset that covers things like news, financial documents, filings, press releases, and even social media, among others. It could potentially revolutionize the dialogue about sentiment analysis and various query-related tasks. This is really only just the start of what's to come for the future of the financial industry.

Suffice it to say, too near is becoming way too near.