In an interview with finews.asia, the co-founder of a Chinese equity boutique talks about how natural language AI chatbots are likely to turn the investment industry upside down.
Qi Wang, born and raised in Shanghai, is a self-described Wall Street veteran. He attended college in the US, graduating with a math and computer science degree from Colby College - and an engineering degree from Dartmouth.
His first job was at Goldman Sachs in New York, and he followed that with stints as an investment director at Elliot Advisors in Hong Kong and as a global tech strategist at UBS in both Hong Kong and New York. He subsequently managed money for state-owned China Everbright Limited, after which he joined MSCI, where he led research into the inclusion of A-shares before co-founding MegaTrust Investments.
In an interview with finews.asia, he looks at the wide-ranging changes expected to take place in the investment industry with the last year’s introduction of natural language processing AI chatbots.
Mr. Wang, how is Chat GTP changing the investment industry?
I would say that prior to Chat GTP, AI was used in finance more quantitatively. Many fund houses were using them for optimization and modeling. With natural language processing, the application of AI increasingly shifts from quantitative to qualitative. If I naively divide the investment world into quantitative and fundamentals, the threat from Chat GTP and similar chatbots is now more on fundamental research.
But how will it change more exactly?
I think for starters we shouldn’t view it just as another tool. It is much more than that and I am thinking deeply as to how this will change our processes.
For example, CHEGG, the US-based textbook rental tech company, recently said Chat GTP would affect its ability to acquire customers and the stock fell 40 percent. The reason I mention this is that it is an education services company that interacts with humans. In essence, we are in the same business. We are all in the services business, and we serve people.
Any industry that has a model to serve and interact with people is going to see a very big impact. There are indications from several sources suggesting that financial services will see the largest impact from AI – after the tech industry of course. More so than retail or healthcare, etc….
I understand. That is the larger picture. But I think many of our readers work in the financial industry and they are interested in the details.
I don’t know yet what the detailed change and individual steps will be. Even though I studied computer science, I can’t say that I fully understand what is going on with AI now. Neither do many of the colleagues that I studied with. But I do believe AI will eat our lunch. What is fundamental research? You read reports, you interact with company management. That is information gathering. Much of that can be taken on by AI chatbots now.
Some skeptics will come out and say that AI can never take over company management. So the core of fundamental research will still be human-to-human interaction. AI will have limits. Machines can’t read between the lines or a person’s body language. A human analyst will always have certain advantages over AI when studying a company run by humans. That is the basic counterargument. However, how sure are you going to be that the CFO and the investor relations (IR) person, or even the CEO, are going to be human in the future? There may very well be companies managed by AI or just algorithms. Does a human analyst have any advantage over AI when analyzing an AI-run company?
I am thinking too far ahead. Some Chinese companies are already using AI for IR, for example, answering basic financial questions from investors online. I think they were forced to do this since there are over 200 million investors in China and you can’t have a human IR team to answer all the questions.
What I am trying to say is, our industry is still tangled up in the old model, thinking of AI as just a new technology that we can use. If you see this as a tool, you are limiting yourself and not seeing the endgame. I don’t see this as just another tool but something that will turn our industry upside down.
What then are the implications for those who work in finance?
When it comes to investing, what we are not doing now is just thinking about the applications. Like I said that is the wrong framework. AI is not going to work nicely with our traditional way of investing. So we are sitting down, taking a step back, trying to learn the technology, thinking about the future, and seeing how we can transform ourselves faster than technology. It won’t be easy.
As a financial industry leader, you should do what I just said above. Not asking your teams to leverage this, or implement that, or how to make a few extra dollars from AI. The big point is to think about how it will shape the future, if there is even a future, and how to survive. To survive, you need to run faster. Simple as that.
What prompts you to say that?
The reason? Let’s forget AI for a second. Take indexing for example. Like AI, indexing is also run by machines, though not as smart. An index is rebalanced four times a year, in some cases just two times a year. Indexing or passive investing now comprises 30-40 percent of the market trading. This means if everything goes into indexing, the market only needs to trade four times a year. The market only needs to open four days a year, perhaps. Scary!
If AI takes over all investing, I think something similar will eventually happen. The market turnover will go down not up. The stock market doesn’t need to open every day. Why? Machines will eliminate a lot of the excessive trading driven by human emotions and biases. Those irrational trades will disappear. Look at this another way. Given enough time, AI will figure out not to act like a nervous day trader but to mimic someone like Warren Buffett. Warren Buffett doesn’t trade that much.
I know there is a strong connection between AI and high-frequency trading. There are firms that are extremely good at this. But if AI takes over the investment world, high-frequency trading will be a dead end. First, there will be no human traders to exploit. Then it will be machines against machines. Trading is not exactly rocket science. So the machine-to-machine battle won’t last for years. Soon it will be over. Then what?
Are you startled by all this?
Yes – in a way. But I don’t know what else I can do, other than to change. With that, I mean myself. My skillsets are in research - fundamental research, trading, and in risk management. All of these can be taken over by AI. Do I have a professional skill that can’t be replaced by machines in the future? I can’t name one.
But, still, you have created a course to raise awareness about AI. Everything can’t be all bad.
The AI course was to raise awareness as to think about self-change. Again, do not just use AI as a tool, but let the technology change you first. I am not sure if I made it clear earlier. I view AI as competition, not a tool. And we need to run faster to beat it.
Are there any optimistic takeaways for the industry and employees?
Yes. The quality of generative AI results largely depends on the quality of the question. A good research analyst is good at asking good questions. I personally go to great lengths preparing the questions before a company meeting, so I can get the most information out of an hour-long meeting, sometimes shared with other analysts. I don’t ask dumb questions. I think this may be one thing that’s still useful for now.
What about Swiss banks? What should they be doing?
I see too many private banks, Swiss or not, still just thinking about AI as another tool. There isn’t much thinking about AI as something disruptive that may also destroy things. One thing I didn’t talk about today is the client side. What do you think of giving one’s estate or inheritance to be managed by an AI agent? There are a lot of good reasons why this is a bad idea today. But you know what, it will happen. Banks need to be ready. The solution is not straightforward.
