ChatGPT can predict stock prices better than humans
ChatGPT is apparently not only capable of composing texts that sound as natural as if they came from a human pen. It now turns out that the chatbot developed by OpenAI is also capable of predicting stock prices.
ChatGPT becomes a trading pro?
Currently, chatbots like ChatGPT are the talk of the town. This is not surprising given the possibilities that such AI solutions open up. For example, the bot can not only calculate complex mathematical calculations, including the calculation path, and design sophisticated source texts. On top of that, it is capable of creating natural-sounding texts without lacking human creativity. It is therefore not surprising that Microsoft has invested the proud sum of 10 billion US dollars in the project. Since developer OpenAI unleashed the tool on the world, more and more chatbot possibilities are opening up. This is mainly a result of interested people experimenting around with artificial intelligence. Alejandro Lopez-Lira, a finance professor at the University of Florida, is one such person. He created an experimental setup in which he had the chatbot predict stock prices.
As a basis for this, he fed ChatGPT headlines from the current press. The bot analyzed these and, as a result, indicated how this might affect the stock price. His result sounds promising. For example, the professor was able to find out that the AI was significantly more reliable than mere chance, which private investors in particular rely on. ChatGPT was surprisingly good at predicting price returns. This news is now shaking up the next industry in the job market. After all, analysts in the financial market are currently paid dearly for their predictions. Should companies be able to simply consult ChatGPT or other chatbots in the future, there would be a gigantic savings potential. According to a report by CNBC, 35 percent of the industry could be threatened.
Chatbots are becoming oracles
It is the efficiency of chatbots in particular that is now likely to cause a stir among some companies in the financial industry. Alejandro Lopez-Lira’s experimental setup already makes this clear. He fed his chatbot in the ChatGPT 3.5 version with more than 50,000 current headlines. The professor then asked whether the share price would develop positively or negatively as a result of the reporting. The results made it clear that AI delivers better results here than the pure “gut feeling” of humans. And this seems to be just the beginning. After all, ChatGPT 4 has already been launched in the meantime, which is supposed to work more reliably and faster. It is therefore not surprising that several hedge funds are already looking with great interest at the experiment of the University of Florida. The finance professor dampens the expectations a little, however. After all, he is concerned that with increasing use of ChatGPT as a stock oracle, its reliability is also likely to decline. We are curious.
Gehört zum Inventar