by Detlef Glow.
To be honest, I have always been a sceptic when it comes to the usage of artificial intelligence (AI) in the finance industry, especially with regard to market reports and in-depth analysis. But since there is so much noise around ChatGPT, I gave it a trial to see if this interface can hold up with my expectations caused by all the headlines about this new AI. To unveil another secret, as I just started to explore the world of artificial intelligence for the write up of articles and reports, I might have not used the best prompts to question the AI.
The question I asked ChatGPT was to write an article on the history of the European ETF industry. Knowing that the library of the AI at that point of time only extended to December 31, 2021, I hadn’t expected any coverage for the year 2022 at all. But I know there are plenty of articles on this topic available via a simple Google search (including this annually updated series by myself) which could be used as easy to access sources for anybody who wants to write an article.
But the results I got were not really impressive. ChatGPT acknowledged that ETFs exist, but the article was rather focused on index funds than on ETFs. That said, the article was not particularly wrong, but very general and had a somewhat incorrect focus. But what bothered me most was the fact that there were no sources referenced to cross check the content. Again, I admit that the prompt I used was definitely not perfect, and some more tries may have brought up better results. But an AI should be able to deal with imperfect prompts and still deliver good results. This is what intelligent humans can do, right?
On the other hand, we know that ChatGPT was able to pass a master exams at Wharton university and may help all kind of students to write their essays in the future.
That said, I am pretty sure that AI—in this case ChatGPT—will be able to generate good articles and analysis with the depth needed to make them meaningful for all industries and in particular, in the financial industry, especially when it comes to “simple” number crunching and the respective commentaries. But I am also pretty sure that AI will miss out the human part of a good analysis, which in a lot of cases can only be added with human interaction such phone calls to ask market participants if they witness the same trends and what they expect to happen from these trends in the future.
Therefore, I think that AI will be a great tool that helps to create basic research and will therefore free up some time for the respective analyst, journalist, student, etc., to add value to their articles, reports, or essays by conducting in-depth research by communicating with other humans.
From my point of view, artificial intelligence and humans will develop a very fruitful partnership in the future, which will add a new level of depth to all kinds of analysis, essays, or reports and will therefore help to increase the knowledge and a better understanding of the topics covered in the respective reports.
This article is for information purposes only and does not constitute any investment advice.
The views expressed are the views of the author, not necessarily those of Refinitiv Lipper or LSEG.