ChatGPT vs. Financial Analysts: Can AI Outperform Human Predictions in the Stock Market?

0:00Today on the podcast, we are going to be talking about can chat GPT. predict the stock market and this is also in addition to a very interesting news headline about someone claiming to use chat gpt to predict winning the lottery numbers and actually winning so today on the podcast we are going to dive into all of that we're going to be talking about what is happening in the industry in finance specifically and are we about to replace financial analysts so all of this comes from a study which was done by university in florida where essentially they put over 50 ,000 news headlines about different companies dating back to October 2021.

0:37They fed them into the chatbot and then they had ChatGPT evaluate whether the news was good, bad, or irrelevant to the company's stock price. What's important here is it's not just saying like is this good or bad news, it's saying you know relevant to the company's stock price. So you know ChatGPT has to have some understanding of not just you know is this positive or negative for the company, but does this actually negative enough to affect the stock price or positive enough to actually make, you know, those sort of impacts? And like, what are the things that affect stock prices? So I think that, in my opinion, is pretty interesting.

1:11In any case, they then use this to analyze and determine whether, you know, they got the responses back from ChatGPT and they had it predict if the company's stock performance was going to improve or be worse on the following day, following one of these news headlines coming out. So in any case, they did this on the study from all of this past data, and the study found that a significant positive correlation between the ChatGPT scores and the actual next-day stock performance was shown. So companies with a higher score from this ChatGPT test, they tended to have better returns than those with a lower score on the day following the article headlines.

1:56And what was really interesting about this is that ChatGPT actually pretty much it outperformed traditional what you would call traditional sentiment analysis methods. So this is already something that's popular in the industry, right? You already have financial analysts doing this kind of thing, and different models and software, yada, yada, that kind of do this. But ChatGPT actually outperformed those traditional methods. And so what's interesting is those traditional methods, they do news headlines, social media, like coverage, how people are responding to them. And they kind of use a bunch of different things to predict it.

2:28But even with that kind of richer data and more real time live data, ChatGPT still did a better job. So the researchers concluded from this, that incorporating advanced language models like ChatGPT into, you know, your investment decision making process is absolutely a good way to, you know, more accurately predict what's actually happening in the stock market. So, you know, this is really interesting. Obviously, they had the disclaimers like don't use ChatGPT for all your stock market, you know, financial investing advice obviously makes sense. But I think it's pretty interesting because obviously, this is showing that these traditional models that we've been using for a long time are not even as powerful as ChatGPT.

3:11And I think that there's a massive opportunity here for financial analysts to use this number one or be replaced, right? Like we're seeing in a lot of different industries. I recently heard from someone in our Facebook group, he had a company that did writing for people and obviously everyone, he lost a lot of clients because we were switching over to just getting ChatGPT to write articles for SEO and that kind of stuff. And so he had to switch his mindset from, you know, I'm not going to let ChatGPT replace me, I'm going to be the person that knows how to use ChatGPT the best so I can stay on top and still offer value.

3:46So I love that mindset shift from him. And I think we're going to see that with every industry, whether that's ChatGPT or any AI or any innovation in any industry, right? Like some people are going to roll over and, you know, think like their career or their industry is over. And the people that are going to win are the ones that use it, leverage it, and become the best at using these AI and these new technological innovations. So anyways, aside from that random tangent. I think all of this is really relevant though, because we have also seen that, you know, we've reported earlier on Bloomberg GPT, which essentially is Bloomberg News that has went and kind of has this massive data set of all of these financial documents, news filings, press releases, social media from a whole bunch of different companies, you know, they've been doing this for a long time.

4:29And they created Bloomberg GPT, which they claim is a model that's going to improve on sentiment analysis, news classification, headline generation, question answering, and a lot of other query-related tasks in the finance space. So it's going to be interesting to see if Bloomberg GPT is as effective as ChatGPT at this. And it's kind of interesting because this isn't just something with Bloomberg. This is a lot of companies around the globe that are doing these kind of using their data sets to train these new models. And it's really interesting. I what I'm really curious about is if these companies are going to be able to compete with chat GPT or if you know chat GPT with its massive data set will have an advantage over these like highly specific models that's something that we're gonna have to watch when these are launched obviously a lot of these things are just announced and not launched right away but we are able to use chat GPT right away so I think this is really interesting and then one other funny bit of news in this kind of chat GPT prediction tangent is that there was recently an article about a guy in Thailand.

5:37His name is Pata HaKorn Boren, and he was on TikTok talking about how he used chat GPT to predict the lottery and won. And it's not really that much money. I think it's about 2000 Taiwanese dollars or whatever their currency is there, which is the equivalent of about$60 US. So not that much money but it is interesting to see what his method was. So he said that he put a few hypothetical questions and some winning numbers from previous years into ChatGPT and asked it to come up with some new you know some some numbers that would help him to win and he was talking to a local news outlet and he said you know he's been using ChadGPT for a while to generate winning lottery things, and he said on several occasions, all fruitful.

6:30I mean, so I'm assuming he's claiming to have done this before, but he really only has proof of the one time. And so, yeah, he says that ChadGPT told him that the winning, that winning, first off, it told him that winning the lottery was about luck, and it reminded him not to be obsessed with this new hobby. It also told him to go outside and do something like exercise, which I think is pretty funny. But it did give him some numbers. He did use them and he won. So I would love to hear what your opinion is on that. If you join our Facebook group, let me know. Is he lucky? Was ChatGPT a genius? Did he strike gold?

7:02Or is the lottery system going to have to revise how they do things? Or is it kind of a coincidence that maybe a bunch of people tried to do that? I don't know. In any case, I think it is, you know, this whole stock market analysis angle is obviously a lot more concrete a lot more data behind it than the you know the lot winning the lottery and so I think it's going to be a really big area that financial analysts are going to have to use or adapt or be replaced because chat gbt is really good at predicting the stock markets next day moves based off of the news and that is something that previously you know we were getting big hedge funds and highly skilled highly trained people to do so just another area that seems to have been automated and has potential for some massive disruption in the future.