PKU Financial Review: Hi Professor Rossi, thank you very much for accepting our interview. We are very impressed by your research related to robo‑advice and personal finance. To start with this interview, could you please summarize the current value of robo‑advice in household wealth management and what remains the biggest challenge?

Professor Rossi: Okay, right now I would say that robot advisors are definitely playing a huge role in democratizing access to financial services and financial advice. Let me just give you an example. Most of the time, in particular in the U.S., you have that a lot of human financial advisors are paid on the basis of the assets under management. So they will charge a fee of the assets that the investor has. What does it mean? It means that if you want to have a human financial advisor, in order for it to make it worth their time, you need to be relatively wealthy. So for those people that don’t have enough level of wealth, they’re cut out completely of these services. There is a ton of individuals that would love to have some level of guidance in how to handle their money, and they’re simply not able to find it. Now, robot advisors are particularly important exactly because they come in and help those people. They are able to provide them with sometimes relatively simple forms of financial advice that enables them to be best in the stock market, be able to have their money grow, and so accumulate wealth over time. Now, the biggest challenges that robot advisors have is that it’s not that trivial to have an individual trust a platform or an app to guide their financial decisions.

PKU Financial Review: Yeah, that’s the second question I want to ask because your research on robo‑advice transforming households pointed out that robo and AI can turn households into rational economic agents, but now many people don’t trust, don’t believe in robo‑advice. For example, in China, many advisory services provided by banks seem somehow rudimentary. So how would you evaluate the current state of these robo‑advice services?

Professor Rossi: Yeah, I think it’s definitely true. In many cases there is this complication right now. In order to provide financial advice in a 360‑degree fashion being extremely effective, you need to have that number one, individuals are able to trust the algorithms, and also you need to have that the algorithms and the platforms are very, very well designed. These two are very difficult to achieve. Number one, because automatically telling someone how much they should save, how much they should invest, what should be the optimal portfolio, how they should set aside money for their college kids, for their down payment for the house – that is very complicated. So right now, at least in the U.S., the solution has been to try to have a hybrid form of advice. Instead of having the algorithm directly interacting with the consumer, in many cases the algorithm is powering a human advisor. What they are doing is they are able to make the advisors more effective and more efficient. Instead of covering – I’m just making an example – instead of having an investment advisor being able to service 200, 300, 400 clients, now because a lot of the tasks like compliance, like portfolio construction, are automated by the algorithm, they can serve many more customers. The role of the human is the one that establishes trust with the client, guides them, and is able to do what we call emotional coaching with the client, trying to understand exactly the financial situation of the family in order to provide the best possible service.

PKU Financial Review: Thank you. And you also proposed a holistic approach covering consumption, savings and debt. So we are wondering, is the current technology mature enough to support this holistic approach, or are we still in the early stages?

Professor Rossi: I think still early. There are a lot of startups that are trying exactly to do it – how would you provide holistic robot advice? Number one, the key challenge is access to the data for the users. In many cases, financial services firms don’t have access to the full picture of the individual or of the household. You may have that if you work with a certain bank or have money with a certain financial advisor, they may see part of your portfolio but not everything. Now there are a lot of companies, in the U.S. in particular, that are trying to be what we call aggregators. What they try to do is accumulate as much information as possible from the clients. Some platforms will ask the individual to link their bank account, their portfolio, their mortgage, their credit cards, and so forth. That will allow the platform to have a 360‑degree view of what the needs of the client are, and then at that stage you can start thinking about optimizing behavior across all these dimensions. Now, is the technology there yet? No. Is there – for example, going back to hybrid forms of advice – a lot of platforms that are enabling financial advisors to do it? The answer is yes. Many financial advisors are using exactly these platforms, and the idea is to cater not only to the individual but to the whole family. Historically, the financial advisor was thinking about servicing the client. Now they are thinking about servicing the family – they will meet with the husband, the wife, the kids, and then they will try to figure out what the needs are for them. This can be on everything: long‑term care insurance, memory care, and everything related to financial planning. Historically, some of these services were limited to what we call family offices – ultra‑high‑net‑worth individuals would have a team of people taking care of estate planning, real estate planning, etc. Now a lot of financial services firms are trying to use technology to make it more and more affordable for every single family to have access to this holistic type of advice.

PKU Financial Review: Thank you. And some research points out that women receive worse financial advice, and your research argued that robo‑advice could either decrease or widen economic inequalities. So under this situation, how would robo‑advice play a role in reducing wealth inequalities?

Professor Rossi: I think it’s a very good point. There are two sides to this question. On the one hand, algorithms in many cases have been vilified in the sense that people have shown that if you train an algorithm on past decisions, and those past decisions of humans were biased, then the algorithm will inherit some of those biases, and so some of the recommendations will be biased. I don’t think this is necessarily true – it is all about the design of the algorithm. You may design an algorithm that automatically corrects for biases. Regarding the second part – whether robot advisors can increase or decrease wealth inequality – I think it really depends on what part of the distribution you are thinking about. In many cases, it is true that robot advisors are able to provide services to people with lower levels of wealth because they are cheaper, so from that perspective they help the part of the population that is not the very wealthy. But there is also – and this is very important in the U.S. – a big chunk of the population doesn’t have any wealth; in fact, they have negative net worth because of student debt, credit card debt, and so forth. So one thing I think we should be working on a little bit more is to try to provide services for individuals to be able to better manage their debt – doing robo‑advice for debt management and trying to help them get out of debt. That would be one solution to make sure that everybody benefits from the introduction of all these new technology tools.

PKU Financial Review: That’s a good idea. And the last question: looking ahead, what do you see as the major innovation of robo‑advice in the next few years?

Professor Rossi: Well, I think that right now, over the last couple of years, one of the biggest challenges that financial services firms have been trying to struggle with is, of course, the introduction of large language models. On the one hand, there is this great desire to introduce direct‑to‑consumers AI tools to increase engagement with the platform and make sure that individuals get exactly the level of personalized advice that they would want. Of course, there are a lot of limitations. Some of these tools are what we say “hallucinate” – they give incorrect advice. This puts the financial services firm at litigation risk, because once the tool has given poor advice and that has been recorded, you are potentially liable. So there is a lot of innovation that is now being introduced in the context of powering advice with some of these technologies – thinking about, for example, meeting summarization between client and advisors, trying to use the network of advisors to understand what are the common concerns that clients have at a specific point in time, what particular economic shocks are particularly important to households. So I think that this field is moving very quickly into trying to incorporate a lot of technology to improve the way financial advice is delivered.

PKU Financial Review: And for Chinese investors, do you have any key advice for them on using AI wealth management tools?

Professor Rossi: No, I think one thing that I highly recommend is that it’s never too early. We are entering a world where longevity is increasing, and people born today are going to be living much, much longer than they think. So I think it’s never too early to start saving. There are a lot of tools nowadays, thankfully, that allow people to save and invest. Now, there are also a lot of competing platforms that engage individuals in, for example, doing a lot of betting on economic outcomes – think about prediction markets. So what I would recommend for the Chinese audience is to definitely try to invest for the long run, have a very well diversified investment portfolio, and start doing it early because compounding is particularly powerful. What you save in your early years is going to pay off tremendously when you retire.

PKU Financial Review: Thank you very much. Many thanks to Professor Rossi. And that’s all for today’s hotspot program. See you next time. Bye bye.

Professor Rossi: Thank you, bye bye.