JJ Hornblass
Hi, everyone. I’m JJ Hornblass and welcome to The Buzz from Bank Automation News. This is our weekly wrap for what’s going on in the industry as of June for 2021.  A hearty thanks to Bank Automation News, advertisers Zeta Tech, and MX for their support, so thank you to them. And Myra Thomas, Loraine Lawson and Jaspreet Kalra. Join me today from the bank automation news editorial team. Hello to all three of you. Let’s start first with some general tech news. A survey of Americans was released today that found that since 2019, smartphone ownership has increased 4% to 85% of Americans, not sure who that remaining 15% are and broadband subscriptions have risen 4% to 77% of all Americans. However, the report found that 30% of those people report internet connectivity issues. President Biden has issued an order banning U.S. investment in 59 Chinese tech companies that aid surveillance substantially expanding on an earlier order issued by President Trump. And the US Department of Justice says that it is elevating investigations of ransomware attacks to a similar priority as terrorism in the wake of the colonial pipeline hack and two bank automation industry news. This week, we reported on the FDI C’s actions to and plans to expand AI usage in banking. We spoke with the FDIC, his new Chief Innovation Officer, Jaspreet tell us a little bit about him, and how the FDIC views AI for banking today.

Jaspreet Kalra
So the Innovation Officer Sultan Meghji is also the inaugural chief innovation officer that the FDIC is appointed. And they’re going to be hosting a lot of banks and fintechs and other interested stakeholders over the next two weeks to sorry, for starting the week of June 14, and to talk about what they’re thinking about when it comes to implementing AI and ml systems for things like cybersecurity, automation of back office processes, credit underwriting, and what’s the sort of takeaway from that conversation was that not only is his his job aimed at how to separate these sort of integrations of new technology into the system, but also look at the FDIC, his own internal operations, and how to sort of modernize those and bring them up to speed with the attack that’s going around in the financial industry. Now, some of the key highlights As for me, according to the conversation were a conversation around AI and bias and credit underwriting also needs to be expanded to look at the existing biases that already exist in the banking system. And run the examples he gave me is that if you say use a prepaid cell phone number, in a banking application versus postpaid cell number, you could end up having different costs associated to accessing certain products and those sort of biases are already baked into the system. So his entire point there is that not only is it about extricating those biases, but also guarding against the fact that AI shouldn’t end up taking new biases into the system. And another thing that I touched upon with him who is in his recent filing in response to the request for information, the federal regulators had put out an AI in financial services UI but noted that an unable global framework at regulating AI B. in countries like Singapore, you are emerging regulations in the US could pose a competitive disadvantage to firms building these systems. And he emphasized the point that irrespective of what happens, the US is going to be sort of like a welcome host to building these systems and AI in financial services could be a real competitive advantage that us firms have going forward.

JJ Hornblass
So Myra, Loraine to take the take the viewpoint of a of the financial institutions and and how does this plan or these positions that are being taken by the FDIC and its new chief CIO? How do they kind of resonate with what you know are the challenges or needs in the banking industry as it relates to AI or Ai, regulation and or leadership.

Myra Thomas
Now, I think what just raised said, you know, as far as UI path has been echoed in the various vendors that we’ve been talking to is that, you know, there’s no clarity. And so as far as the technologies that and automation that they’re developing product launches that we’re working on there, you know, they’re concerned, obviously, because there is no clear, clear path for them. So, you know, I think that’s one major consideration. As far as the bags, I think rain had some interesting tidbits that she got, as far as her feature that she’s working on on ethical AI as well, which talked a little bit about, you know, the unintended consequences. And I’m sure it’s such Forgive me, I think some of the banks will be concerned about the unintended consequences of AI. And so maybe that will make them a little bit more cautious about expanding cash, initially, in large quantities, you know, on this technology, simply because of this unintended consequences. So I think they’re going to have to figure out, you know, there’s a lot of providers out there who are starting to say that they, you know, are in the AI focus, you know, Rome, I think the Chief Innovation Officer at the FDIC came from ones themselves, and he created, um, we’re headed out. So for banks, I think it’s a big question mark, you know, at least right now, and it’ll be interesting to see when how much they actually decided to devote as far as finance funding towards AI related automation.

JJ Hornblass
Loraine, how does this jive with the ethical AI issues or sort of the the needs of the marketplace? For for leadership guidance or enforcement,

Loraine Lawson
currently think banks are looking to regulators, I think that’s a big part of why they’re being slow to adopt AI in areas like credit decisioning. Although that is automated, they don’t necessarily use AI to do it in an automated way. So I was talking to Stephen Thomas, he’s executive director, Director at the analytics and AI ecosystem in the Smith School of Business at Queen’s University that’s in Ontario, Canada. And he had actually he actually served as being about this tie has talked to them, primarily in North America. But what he has found is that vapes are researching how to use AI and credit decisioning, but they still haven’t implemented anything. Now, I do think that that guidance is underway. There have been a number of legislation active activities around this recently, and may 27. The Senator Edward J. Markey and Congresswoman Doris met Sue Matsui, sorry, introduced a algorithmic justice and online platform Transparency Act in Congress to prevent harmful algorithms increased transparency into website content application and moderation practices, and also to create a commission that would be across the government and would investigate discriminatory algorithmic processes in throughout the economy. So I definitely think we’re gonna see more activity around this in the next couple of years.

JJ Hornblass
One of the things that striking jaspreet about what you learn from the FDIC there, you know, that he clearly mentioned, the need for coordination with other jurisdictions. There, he did not mention, at least I don’t think he did not mention coordination with other US regulatory bodies, we have three at least three other financial institution enforcement bodies in the United States. And as I understand it, I think that the FDIC is the only one with a Chief Innovation Officer at this point. Is there a sense for what kind of a mo, you know, memorandum of understanding that may be inked between the regulators that might help with some of this kind of standardization? Is there a focus internally, or is it all about, you know, extra us activities.

Jaspreet Kalra
So a couple of points. So he did talk a little bit about how a lot of regulatory agencies do work on this and they do internally coordinate. I mean, if you go back to the RFI, that the agencies that issued, all five of them issued it jointly. So they’re sort of in the process of learning more about Right now and his overall expectation was that if rules do come out on AI, eventually, it’s unlikely that they would be overtly negative about using AI United States. On the other hand, I don’t think in doing about the internal upgrades, there’s a lot of coordination about how internal upgrades are being pushed out, that’s more of a agency specific thing about how each agency’s trying to build onto those things. Like, for instance, he talked about the FDIC, having used AI systems for about five years now for some processes. But on the other hand, they still have older technology that they’re looking to upgrade. But one of the things he also emphasized was there, a lot of fixation tends to happen around credit underwriting and AI use. And his entire point was that credit underwriting is a very small part of AI used in financial services, a lot of it is managing large scale technical infrastructure, some of it is also AML Compliance that can help banks detect patterns that can otherwise manually take literally days to calculate or look at. So I think it’s a much broader conversation. And it’s a lot about, which is why these office hours are sort of their first move on this to just get people into the door, and talk to them as much as they can about to learn what they’re sort of thinking about. And so far, there’s been good interest expressed, like the number he gave me was over 100 entities want to come in and talk to the FDIC about what they’re thinking about. So it could be a slow moving train, but I do think there is going to be coordination among federal agencies, when they think about proposing rules on this.

JJ Hornblass
Is it my around the rain? I mean, is it is it on? Is it the case that that there is this fixation on credit underwriting? Or, you know, is that accurate? Or and maybe, is it founded considering the the dollar size of that? of of credit in the United States?

Loraine Lawson
I mean, I was gonna say, you go ahead, sorry.

Myra Thomas
No, I, you know, I understand the concern, you know, given history and all sorts of things, and redlining, what have you? So, yes, the focus would seem to be on the impact credit underwriting. So, that does make sense. But obviously, there are many, many more ways where AI can start, you know, I could cause a problem in banking. I think rain pulled up some that initially even involved marketing, you know, products, but, you know, as with any new technology, it’s always the implementation, the data that goes in that, you know, automatically may change the outcome. And that much of that can be intentional or unintentional. You know, the human factor does matter. So, you know, it’s, it’s a difficult proposition. But yes, the focus on credit underwriting does make sense. I mean, obviously, you know, they’re always big players, and big banks now trying to figure out ways to, you know, get around traditional ways to, to underwrite individuals, and they assume that AI will help them figure out ways around, you know, traditional ways, more traditional ways that sort of exclude the underbanked or unbanked, by finding new metrics way, ways that people operate by using AI. But, you know, little is known about how that’s actually going to work out. Now, the fact that, you know, the major banks from JP Morgan, etc, will be in it. Obviously, they understand that it’s a valuable marketplace to be in

JJ Hornblass
the rain, what’s the scope of the effort to automate credit decisioning using AI today? I mean, how, how much of an effort is being put into integrating AI into credit? decisioning?

Loraine Lawson
No, there’s actually a vendor that does this, and he is getting clients. But I think that’s just not known, because the banks are very, very cagey about talking about it at this point, because it is so much is unknown. Myra mentioned the marketing efforts, to be fair, that that is that should be looked at as well, because banks are using AI Now, we know to create personalized credit offers to people and so you know, personalized rates or personalized offers, and they’re using data that could potentially be biased because it you know, historical data, unfortunately and codifies allows societal bias. And that’s, that’s the challenge in AI is looking at, you know, really, really looking at your data sources. And right now, the problem is there aren’t a lot of best practices about how to fix the data like that. You Think you would just pull out all the demographics, but it doesn’t really work like that. Because AI is so good at pattern recognition, it can pick up on subtle things that would would reveal you, for instance, as a white male of a certain age like me, your buying habits could be used to do that. So these, these things get codified in ways in the data that humans don’t necessarily recognize. And then, of course, as Myra points out, it’s a little naive to think that some of this wouldn’t be intentional, at some point, maybe not on the financial institutions part, but on, you know, other areas where people would use it. Yeah.

JJ Hornblass
Right. So the upshot is the FDIC has to work cut out for it. What else do we have planned for our coverage next week?

Myra Thomas
Next week, we’re working on a number of things. I know that in the hopper, we have a story that will clarify, you know, the differences between what process mining and process discovery really named, since there’s been quite a bit of financial activity in that realm, as far as venture capital investment, and also, I have another company that’s quite large, it’s in our funding round up at the time, about 50% of that market, as well, as I believe, if you guys want to talk a little bit about what you’re working on as well. I’m also working on a story involving cross border AI, utilize to fight against anti Mondor money laundering. And so there’s a lot of activity on the cross border payments side. And so obviously, the next stage for that next step for that is better, you know, better intelligence and better automation as far as AML. So I’m working on a story on that as well. Yeah, another thing involving a Walmart deal that I can’t get mentioned. That will come out on Tuesday.

JJ Hornblass
Great. Well, I urge everyone to visit us at Bank automation news.com. And to follow us on Twitter, and LinkedIn, thank you to buy Lorraine and jasprit for your insights. We will see you next time on the bus. Thank you very much.

JJ Hornblass
Hi, everyone. I’m JJ Hornblass and welcome to The Buzz from Bank Automation News. This is our weekly wrap for what’s going on in the industry as of June for 2021.  A hearty thanks to Bank Automation News, advertisers Zeta Tech, and MX for their support, so thank you to them. And Myra Thomas, Loraine Lawson and Jaspreet Kalra. Join me today from the bank automation news editorial team. Hello to all three of you. Let’s start first with some general tech news. A survey of Americans was released today that found that since 2019, smartphone ownership has increased 4% to 85% of Americans, not sure who that remaining 15% are and broadband subscriptions have risen 4% to 77% of all Americans. However, the report found that 30% of those people report internet connectivity issues. President Biden has issued an order banning U.S. investment in 59 Chinese tech companies that aid surveillance substantially expanding on an earlier order issued by President Trump. And the US Department of Justice says that it is elevating investigations of ransomware attacks to a similar priority as terrorism in the wake of the colonial pipeline hack and two bank automation industry news. This week, we reported on the FDI C’s actions to and plans to expand AI usage in banking. We spoke with the FDIC, his new Chief Innovation Officer, Jaspreet tell us a little bit about him, and how the FDIC views AI for banking today.

Jaspreet Kalra
So the Innovation Officer Sultan Meghji is also the inaugural chief innovation officer that the FDIC is appointed. And they’re going to be hosting a lot of banks and fintechs and other interested stakeholders over the next two weeks to sorry, for starting the week of June 14, and to talk about what they’re thinking about when it comes to implementing AI and ml systems for things like cybersecurity, automation of back office processes, credit underwriting, and what’s the sort of takeaway from that conversation was that not only is his his job aimed at how to separate these sort of integrations of new technology into the system, but also look at the FDIC, his own internal operations, and how to sort of modernize those and bring them up to speed with the attack that’s going around in the financial industry. Now, some of the key highlights As for me, according to the conversation were a conversation around AI and bias and credit underwriting also needs to be expanded to look at the existing biases that already exist in the banking system. And run the examples he gave me is that if you say use a prepaid cell phone number, in a banking application versus postpaid cell number, you could end up having different costs associated to accessing certain products and those sort of biases are already baked into the system. So his entire point there is that not only is it about extricating those biases, but also guarding against the fact that AI shouldn’t end up taking new biases into the system. And another thing that I touched upon with him who is in his recent filing in response to the request for information, the federal regulators had put out an AI in financial services UI but noted that an unable global framework at regulating AI B. in countries like Singapore, you are emerging regulations in the US could pose a competitive disadvantage to firms building these systems. And he emphasized the point that irrespective of what happens, the US is going to be sort of like a welcome host to building these systems and AI in financial services could be a real competitive advantage that us firms have going forward.

JJ Hornblass
So Myra, Loraine to take the take the viewpoint of a of the financial institutions and and how does this plan or these positions that are being taken by the FDIC and its new chief CIO? How do they kind of resonate with what you know are the challenges or needs in the banking industry as it relates to AI or Ai, regulation and or leadership.

Myra Thomas
Now, I think what just raised said, you know, as far as UI path has been echoed in the various vendors that we’ve been talking to is that, you know, there’s no clarity. And so as far as the technologies that and automation that they’re developing product launches that we’re working on there, you know, they’re concerned, obviously, because there is no clear, clear path for them. So, you know, I think that’s one major consideration. As far as the bags, I think rain had some interesting tidbits that she got, as far as her feature that she’s working on on ethical AI as well, which talked a little bit about, you know, the unintended consequences. And I’m sure it’s such Forgive me, I think some of the banks will be concerned about the unintended consequences of AI. And so maybe that will make them a little bit more cautious about expanding cash, initially, in large quantities, you know, on this technology, simply because of this unintended consequences. So I think they’re going to have to figure out, you know, there’s a lot of providers out there who are starting to say that they, you know, are in the AI focus, you know, Rome, I think the Chief Innovation Officer at the FDIC came from ones themselves, and he created, um, we’re headed out. So for banks, I think it’s a big question mark, you know, at least right now, and it’ll be interesting to see when how much they actually decided to devote as far as finance funding towards AI related automation.

JJ Hornblass
Loraine, how does this jive with the ethical AI issues or sort of the the needs of the marketplace? For for leadership guidance or enforcement,

Loraine Lawson
currently think banks are looking to regulators, I think that’s a big part of why they’re being slow to adopt AI in areas like credit decisioning. Although that is automated, they don’t necessarily use AI to do it in an automated way. So I was talking to Stephen Thomas, he’s executive director, Director at the analytics and AI ecosystem in the Smith School of Business at Queen’s University that’s in Ontario, Canada. And he had actually he actually served as being about this tie has talked to them, primarily in North America. But what he has found is that vapes are researching how to use AI and credit decisioning, but they still haven’t implemented anything. Now, I do think that that guidance is underway. There have been a number of legislation active activities around this recently, and may 27. The Senator Edward J. Markey and Congresswoman Doris met Sue Matsui, sorry, introduced a algorithmic justice and online platform Transparency Act in Congress to prevent harmful algorithms increased transparency into website content application and moderation practices, and also to create a commission that would be across the government and would investigate discriminatory algorithmic processes in throughout the economy. So I definitely think we’re gonna see more activity around this in the next couple of years.

JJ Hornblass
One of the things that striking jaspreet about what you learn from the FDIC there, you know, that he clearly mentioned, the need for coordination with other jurisdictions. There, he did not mention, at least I don’t think he did not mention coordination with other US regulatory bodies, we have three at least three other financial institution enforcement bodies in the United States. And as I understand it, I think that the FDIC is the only one with a Chief Innovation Officer at this point. Is there a sense for what kind of a mo, you know, memorandum of understanding that may be inked between the regulators that might help with some of this kind of standardization? Is there a focus internally, or is it all about, you know, extra us activities.

Jaspreet Kalra
So a couple of points. So he did talk a little bit about how a lot of regulatory agencies do work on this and they do internally coordinate. I mean, if you go back to the RFI, that the agencies that issued, all five of them issued it jointly. So they’re sort of in the process of learning more about Right now and his overall expectation was that if rules do come out on AI, eventually, it’s unlikely that they would be overtly negative about using AI United States. On the other hand, I don’t think in doing about the internal upgrades, there’s a lot of coordination about how internal upgrades are being pushed out, that’s more of a agency specific thing about how each agency’s trying to build onto those things. Like, for instance, he talked about the FDIC, having used AI systems for about five years now for some processes. But on the other hand, they still have older technology that they’re looking to upgrade. But one of the things he also emphasized was there, a lot of fixation tends to happen around credit underwriting and AI use. And his entire point was that credit underwriting is a very small part of AI used in financial services, a lot of it is managing large scale technical infrastructure, some of it is also AML Compliance that can help banks detect patterns that can otherwise manually take literally days to calculate or look at. So I think it’s a much broader conversation. And it’s a lot about, which is why these office hours are sort of their first move on this to just get people into the door, and talk to them as much as they can about to learn what they’re sort of thinking about. And so far, there’s been good interest expressed, like the number he gave me was over 100 entities want to come in and talk to the FDIC about what they’re thinking about. So it could be a slow moving train, but I do think there is going to be coordination among federal agencies, when they think about proposing rules on this.

JJ Hornblass
Is it my around the rain? I mean, is it is it on? Is it the case that that there is this fixation on credit underwriting? Or, you know, is that accurate? Or and maybe, is it founded considering the the dollar size of that? of of credit in the United States?

Loraine Lawson
I mean, I was gonna say, you go ahead, sorry.

Myra Thomas
No, I, you know, I understand the concern, you know, given history and all sorts of things, and redlining, what have you? So, yes, the focus would seem to be on the impact credit underwriting. So, that does make sense. But obviously, there are many, many more ways where AI can start, you know, I could cause a problem in banking. I think rain pulled up some that initially even involved marketing, you know, products, but, you know, as with any new technology, it’s always the implementation, the data that goes in that, you know, automatically may change the outcome. And that much of that can be intentional or unintentional. You know, the human factor does matter. So, you know, it’s, it’s a difficult proposition. But yes, the focus on credit underwriting does make sense. I mean, obviously, you know, they’re always big players, and big banks now trying to figure out ways to, you know, get around traditional ways to, to underwrite individuals, and they assume that AI will help them figure out ways around, you know, traditional ways, more traditional ways that sort of exclude the underbanked or unbanked, by finding new metrics way, ways that people operate by using AI. But, you know, little is known about how that’s actually going to work out. Now, the fact that, you know, the major banks from JP Morgan, etc, will be in it. Obviously, they understand that it’s a valuable marketplace to be in

JJ Hornblass
the rain, what’s the scope of the effort to automate credit decisioning using AI today? I mean, how, how much of an effort is being put into integrating AI into credit? decisioning?

Loraine Lawson
No, there’s actually a vendor that does this, and he is getting clients. But I think that’s just not known, because the banks are very, very cagey about talking about it at this point, because it is so much is unknown. Myra mentioned the marketing efforts, to be fair, that that is that should be looked at as well, because banks are using AI Now, we know to create personalized credit offers to people and so you know, personalized rates or personalized offers, and they’re using data that could potentially be biased because it you know, historical data, unfortunately and codifies allows societal bias. And that’s, that’s the challenge in AI is looking at, you know, really, really looking at your data sources. And right now, the problem is there aren’t a lot of best practices about how to fix the data like that. You Think you would just pull out all the demographics, but it doesn’t really work like that. Because AI is so good at pattern recognition, it can pick up on subtle things that would would reveal you, for instance, as a white male of a certain age like me, your buying habits could be used to do that. So these, these things get codified in ways in the data that humans don’t necessarily recognize. And then, of course, as Myra points out, it’s a little naive to think that some of this wouldn’t be intentional, at some point, maybe not on the financial institutions part, but on, you know, other areas where people would use it. Yeah.

JJ Hornblass
Right. So the upshot is the FDIC has to work cut out for it. What else do we have planned for our coverage next week?

Myra Thomas
Next week, we’re working on a number of things. I know that in the hopper, we have a story that will clarify, you know, the differences between what process mining and process discovery really named, since there’s been quite a bit of financial activity in that realm, as far as venture capital investment, and also, I have another company that’s quite large, it’s in our funding round up at the time, about 50% of that market, as well, as I believe, if you guys want to talk a little bit about what you’re working on as well. I’m also working on a story involving cross border AI, utilize to fight against anti Mondor money laundering. And so there’s a lot of activity on the cross border payments side. And so obviously, the next stage for that next step for that is better, you know, better intelligence and better automation as far as AML. So I’m working on a story on that as well. Yeah, another thing involving a Walmart deal that I can’t get mentioned. That will come out on Tuesday.

JJ Hornblass
Great. Well, I urge everyone to visit us at Bank automation news.com. And to follow us on Twitter, and LinkedIn, thank you to buy Lorraine and jasprit for your insights. We will see you next time on the bus. Thank you very much.