00:00:00
Well, the finance sector is taking a much closer look at how technologies are going to be playing a potential role in
00:00:06
their operations in the years ahead. Things like artificial intelligence, cyber risk are more on the agenda of
00:00:13
many experts in these areas. And so the Wharton School is going to be hosting a conference this week about some of those
00:00:19
potential concerns. Wharton finance professor Itai Goldstein joins us to talk about the conference and as well
00:00:26
some of these issues and why the conversations are more important in this day and age. Great to see you again,
00:00:32
Itai. Thanks for coming up. Great to see you, Dan. So, let's start about just the the the
00:00:37
element of AI in your area in finance and the potential impact you think that it might have.
00:00:45
Yeah, great question. So, AI is now a very hot topic. I think hot topic across different fields and certainly has a big
00:00:52
effect on finance. If you ask where finance is going to be affected, it's probably going to be affected across the
00:00:58
board. So, different areas of finance will be affected. I don't know that there is an area of finance that will
00:01:04
not be affected. So, it's it's really a systematic effect across the board.
00:01:09
And you know, when we put together this conference in collaboration with the International Monetary Fund, the IMF, um
00:01:16
we really tried to cover all of it. So, there will be discussion on how it affects financial markets. There will be
00:01:24
discussion on how it affects firms. And when it comes to firms, you know, those that are probably more exposed are those
00:01:32
that have labor that is working on tasks that are potentially replaceable by AI. So, all
00:01:40
this will be affected and will come up at the at the conference. And so part of this I understand is out
00:01:46
of some research that you have done looking at the component of AI and at times how it could potentially impact
00:01:53
something like trading. Yes, absolutely. So, we've done work together with Winston Du, who is a
00:02:01
in the finance department with me, and Yan Ji, who is in the Hong Kong uh uh University of Science and Technology.
00:02:08
Um basically, what we did was to think about how AI traders are going to affect the equilibrium in financial markets. Uh
00:02:17
you know, everyone is talking about AI algorithms replacing humans uh in uh trading. Um and one thing that is nice
00:02:25
about AI is that you can actually go out and check uh by just doing the experiment. Uh so, you know, when you're
00:02:32
thinking about humans, uh then if you want to get an answer of how will they behave in this situation,
00:02:39
you can either do an experiment, but doing an experiment with humans is not that easy. You have to uh recruit 20
00:02:45
students, put them in a lab, give them instructions, see what they do, and there are all these concerns about who
00:02:51
is going to participate in these experiments, and whether it mimics the real world environment or not. With AI,
00:02:57
you can actually do the experiment fairly easily because there are no humans involved. So, basically, what you
00:03:01
do is you just program these AI uh traders, uh you program the financial market environment, and then you let
00:03:08
them act, and you see what happens. So, you know, the AIs that we are looking at
00:03:12
in this paper are the reinforcement learning uh AI. Uh you know, the sub uh category of that is known as Q-learning,
00:03:21
which uh is very prominent in in the industry. Basically, what they do is uh they are just trying to maximize their
00:03:29
payoff, and they don't know anything about the environment, who they're trading against.
00:03:34
Right. Uh the only thing they know is uh they have a set of options. Uh every round of trade, they pick one
00:03:40
option, and then they update, you know, uh this was the state of the world, this
00:03:45
is what I did, this is what I got. And they have these huge metrics of state of the world, action that I took, and then
00:03:51
they keep updating, what do I get for every combination of state of the world and action that I took. And and they
00:03:57
kind of learn as they go. So, over time they become better and better, and they know how to choose the right actions.
00:04:04
Now, you know, the system is such that they have to experiment, so occasionally they
00:04:09
would try something new and see if if it works, and then they update, but eventually they converge to what they
00:04:14
learn is is the best outcome. So, we did that. We run We ran this experiment, and
00:04:19
we found very interesting results that in some cases, and we identified those cases, they converge on collusive
00:04:28
behavior. So, you know, what you would like is traders in financial markets not to collude. You know, when they get
00:04:33
information, they just trade on the information. The information immediately shows up in the price, and then the
00:04:38
system is very efficient. A collusive behavior is they get the information, but they don't trade so quickly on it
00:04:44
because they realize that there is bigger long-term gain if they're not trading very aggressively. Um, and you
00:04:51
know, there are all sorts of theories about when humans will collude and so on. What is interesting is these very
00:04:56
basic algorithms that I just described to you end up in what looks like a collusive behavior, where they just
00:05:03
don't trade very aggressively, and they end up with higher profits in the long
00:05:07
run as a result. Right. Should we be surprised that there is this element of collusion potentially
00:05:12
involved in AI? Because I think most people would think that that term really associates with how human beings act.
00:05:22
Right, yes. Um, so I I think that's a that's a great question. You know, why do we see AIs
00:05:29
ending up in this more sophisticated behavior that we tend to expect from humans.
00:05:35
Um, and people have looked at AI collusion in other contexts. So, there are sort of simple experiments that have
00:05:42
been done in much simpler context. Um, and people have seen this kind of behavior emerging. I think what was more
00:05:50
challenging and really the motivation for our paper is to see whether it can also emerge in a financial market
00:05:57
environment because a financial market environment is much more complex. Right. Uh, you have noise traders, right? You
00:06:03
have all these noise shocks hitting the trading process and affecting price. So, collusion could
00:06:09
become more difficult. You have a market maker. So, some someone out there is observing all uh, the trades and decides
00:06:16
on the price and sort of acts rationally. So, when you let all this interact, the question is whether
00:06:21
collusion still still emerges. And and we identified two types of collusion. One is sort of based on a price trigger
00:06:29
punishment if you want. Right. So, the idea is that if I am an AI trader and I see the price going beyond
00:06:36
a bound that I expected, then I realize that, you know, maybe I should also act aggressively. And and this punishment is
00:06:43
what maintains the collusive behavior in equilibrium. Uh, we call that artificial
00:06:48
intelligence in the sense that they are acting intelligently, but this is really
00:06:53
artificial. Uh, but then there is also another type of collusion which we coined the term, you know, artificial
00:06:59
stupidity, which is that uh, they really end up colluding because they don't
00:07:04
realize that if they deviate from collusion, they can actually make a higher profit in in the short term. So,
00:07:09
they just got used to this kind of less aggressive behavior and that's what they
00:07:13
end up doing. So, the importance of doing the conference right now and you obviously
00:07:18
you talked about the partnership with the IMF. Uh, I guess we're coming to a kind of a a
00:07:25
critical mass point here on a lot of these issues around AI and obviously things like cybersecurity risk and how
00:07:31
they can all factor in, correct? Yes, absolutely. So, you know, we started the collaboration with the IMF
00:07:37
last year. Uh, so this is a collaboration between Wharton Initiative on Financial Policy and Regulation that
00:07:44
I'm uh, the director of and we're doing this conference with the IMF. Last year we did a collaboration,
00:07:50
they wrote the global financial stability report on non-bank financial fragility.
00:07:55
I worked with them on that and we said, you know, let's just do this conference
00:07:59
to highlight the issues that are coming around non-bank financial stability. Conference was a big success and we
00:08:04
said, you know, we can think about it as an annual event. Every year the IMF is producing this global financial
00:08:10
stability report and they identify new issues that are on the agenda of financial stability and we can do the
00:08:17
conference around that and this year it happened to be cybersecurity and AI and how they affect financial stability. So
00:08:23
they wrote the report on that. I think it's getting worldwide attention and and that is a good opportunity for
00:08:30
us to team again with them and do the the conference on that. So this these are certainly issues that are high on
00:08:36
the agenda of policy makers. I I we talk so much I I think in our perspective about what's going on here
00:08:43
in the United States, but from a global perspective on the IMF, these are issues
00:08:47
that they are dealing with and have to be focused on on a daily basis. Focused on
00:08:54
different financial systems all across the globe. Yes, absolutely. So that is the mandate
00:08:59
of the IMF. You know, the IMF is kind of an international organization who's
00:09:03
trying to coordinate financial policy, macroeconomic policy around the world. I mean that their
00:09:10
direct activities are basically giving money to countries in trouble, but they also follow countries, follow emerging
00:09:17
risks and are trying to provide advice and and guidance as to what policies should be undertaken. So AI is certainly
00:09:24
a global issue. I mean there there there are no borders to it. Is there while there's a concern about
00:09:31
the level of risk, there obviously has to be a level of optimism about what AI could bring to the development of
00:09:39
financial systems as we move forward. Yes, I think that is absolutely true. You you know, I mean when we think about
00:09:46
AI, I think we have to start from the opportunity. Because AI was not developed to destroy the world or make
00:09:55
the world unstable. AI was developed because there is an opportunity to make things better, to
00:10:03
have things work more efficiently. And we see that I think across the board that AI is very powerful and can do
00:10:11
things very effectively. And and you know, when it comes to financial markets, yes, the idea that
00:10:18
there will be this algorithm out there that can process all this information very quickly and trade certainly can
00:10:24
lead to improvement. But at the same time as we see that developing, I think it's critical to keep an eye on the
00:10:31
emerging risks. And you know, what one risk that everyone highlights when it comes to AI
00:10:37
is that AI is just going to replace humans, maybe cause massive unemployment, maybe when they take over
00:10:45
from humans, they start controlling the system and then who knows where we end up. So this is a major risk. You know,
00:10:51
the thing I talked to you about is is not as extreme as that, but also something to take a look at.
00:10:57
You know, if you have all these AI algorithms out there that are running the trades in financial markets, are
00:11:03
they going to end up without directly communicating with each other? Are they just going to end up colluding and
00:11:09
limiting competition? So that is another form of risk that we have to think about.
00:11:13
What what do you hope that that people will take away from from the conference? So I I think it's an opportunity for
00:11:20
people to come together and talk about issues that are now on top of the agenda.
00:11:27
It happens to be AI and cybersecurity and how they affect financial markets. I think it's really about understanding
00:11:34
what new dimensions of risk are there that that we have to to look at. Um the the conference features six papers on on
00:11:43
these topics, AI and and cybersecurity, but there are also panels. Uh there is an academic panel, there is a policy
00:11:49
panel, so people are going to talk about what they think are the main issues right now and what we should uh look uh
00:11:56
look for. I I think this interaction between academics, policy makers, and uh people working in the industry is very
00:12:03
important. It's really uh an opportunity to exchange ideas and get informed about
00:12:08
what is coming uh on these dimensions. Great to see you again. Yeah, good to see you.
00:12:13
Thank you. Itai Goldstein, Wharton finance professor.