Menu
Banking Exchange Magazine Logo
Menu

First Installment: SAS Executive Stu Bradley Discusses 2024 Anti-Fraud Report and Its Findings

First of a four-part interview that will be displayed this week on Banking Exchange

  • |
  • Written by  Banking Exchange staff
 
 
First Installment: SAS Executive Stu Bradley Discusses 2024 Anti-Fraud Report and Its Findings

Stu Bradley, Senior Vice President of Risk, Fraud and Compliance Solutions at SAS, with Erik Vander Kolk of Banking Exchange. This is the first of a four-part interview that will be displayed this week on Banking Exchange.

BE Twitter Logo large black 45Erik Vander Kolk
Stu, Thank you for spending time here with Banking Exchange.
One of the things I like about SAS in general is just that SAS goes beyond the financial sector in terms of digital technology, yet this global anti-fraud technology study by the ACFE and SAS shows that the financial sector is actually leading many industries in terms of early adoption of AI and machine learning for fighting fraud and financial crimes.

BradleyStu Bradley
Financial services leads in many regards compared to what we see in other industries — sometimes years ahead.
And you look at how large government agencies are now adopting the similar digitalization of payments for things like social benefits and the automated filing of tax returns, to use another civilian example. They start to see the same types of fraud and risk issues that are seen in other industries, so there is a lot of cross pollination, and financial services often leads the way, both in adopting digital capabilities and using technology to combat the threats that inevitably arise.

BE Twitter Logo large black 45Erik Vander Kolk
Anti-fraud programs use of AI and machine learning will almost triple by the end of next year, according to the anti-fraud pros who participated in the survey. But it seems that adoption has historically lagged respondents’ expectations since the ACFE and SAS debuted this research study in 2019?

BradleyStu Bradley
One in four respondents from banking and financial services organizations indicated that they're using AI or machine learning to fight fraud, and another 37% expect to do so within the next one or two years.
These anti-fraud professional are leaders in comparison to their counterparts surveyed from other industries. Even so, what was really profound to learn is how much AI and machine learning adoption has consistently lagged respondents’ expectations. The ACFE and SAS ran this survey in 2019 and again in 2022, and the adoption of AI for fraud detection across industries has only grown 5% since 2019. That figure falls far short of what the anticipated adoption rates were at the time.

BE Twitter Logo large black 45Erik Vander Kolk
Why do you think adoption has lagged expectations?

BradleyStu Bradley
I think the AI adoption gap we’re seeing speaks to a few things. First and foremost, the complexity of AI adoption. There's no doubt that organizations are seeing tremendous promise in adopting AI, but there are complexities they need to address to ensure they get production value from it.

Secondly, there’s not a single definition of AI. AI can help across a wide range of different use cases, so I think the study findings highlight some of the differences. Too many organizations are rushing to adopt artificial intelligence without aligning to very specific business use cases that have supporting value propositions. This results in a lot of really great ideas that ultimately become science projects, because they're not ultimately getting that production value from their AI initiatives. A lot of it becomes waste. With the cost pressures that every industry is under, organizations simply don't have time or money for wasted efforts.

Another issue is that many organizations lack a true understanding of their own analytic maturity, particularly in reference to artificial intelligence. Things like, do we truly understand our data, the quality of that data, how we source the data? What about our model risk management practices? How do we ultimately govern the models within this institution in a fair and accurate way? How do we actually operationalize AI? So, beyond using AI to help facilitate providing better information or helping make more confident decisions, how does that fit into the overall operational paradigm of an organization?

Too many times these considerations are afterthoughts, and they need to be forethoughts, such that they can help select the right use cases where they can deliver value.

Lastly, I think organizations need to greatly mature in their thinking about the downstream impacts and consequences of their AI deployment. This will allow them to establish better guardrails for their processes and also in defining how they insert humans into the loop from an overall engagement standpoint.

BE Twitter Logo large black 45Erik Vander Kolk
Where has AI already shown a positive impact in financial services?

BradleyStu Bradley
We are seeing the greatest success in the financial services industry from an AI adoption perspective in efforts to drive efficiency into what we typically see as very manual processes. Think about automating the mundane tasks and allowing employees to focus more of their time on the value-added tasks at hand. For example, take a credit origination process. You've got a loan officer that spends a great percentage of his or her time collecting documentation, doing research on the individual, and pulling that together into a loan package.

What if we could automate the collection of that information and the presentation of that information into a very concise and structured form, such that that loan officer could spend the majority of their time actually evaluating whether or not they should originate this loan and how they might want to price it? Not only is that going to drive better efficiency, it's going to be a much more rewarding experience for the employee. And one of the greatest byproducts is it will streamline the overall origination process and greatly enhance the customer experience.

BE Twitter Logo large black 45Erik Vander Kolk
You mentioned AI guardrails. Can you give me an idea of what those guardrails should be or how a bank should go through that process? How do you assure using the data for the right purposes and what are two or three things the banks should do?

BradleyStu Bradley
Multiple sectors can see great gains from using artificial intelligence. I gave you a credit risk example. But whether it's better fraud detection or a customer engagement or marketing use case, across the board AI can help banks deliver better customer experience. We've done a lot of research in this area. It's really about fostering customer trust.

SAS has established six tenets of responsible innovation: human centricity; transparency; robustness in your guardrails; privacy and security; inclusivity; and then ultimately, accountability for mitigating potential adverse impacts when you're leveraging artificial intelligence. Ethical, responsible innovation is critical when using AI. If financial institutions can set similar tenets of their own and follow those tenets, I think they have a fighting chance at successful AI adoption.

It's not just the business that needs to be on board with this. You need to understand across the board, from the CIO or Chief data officer perspective — how are you leveraging the data, the efficacy of that data, the robustness of that data — to the business use behind it, to engaging the model governance team to ensure that you're following those principles all the way through, to audit and compliance and making sure that when you’re rolling out use cases that you've aligned those resources to a common vision and how to extract value from it.

I talk a lot about delivering these tenets through what we call the data and AI lifecycle, and it starts with data. SAS has made a pretty significant investment in data ethics. If we get this right, you could scale financial services to serve the underserved communities, or you could provide more healthcare services to those in need while eliminating fraud, waste and abuse out of the system. These are the types of things that we, the collective “we,” could do if we actually get this right. So yes, there's risk, but I think industry leaders are starting to think about what we must collectively do to ensure that we're not only managing the risk but getting it right at an industry level. If we think about AI in that AI, the potential can really be something substantial, not only for an industry, but for society at large.


Be sure to read every installment in our four-part interview series:

SAS Executive Stu Bradley Discusses 2024 Anti-Fraud Report and Its Findings


 

back to top

Sections

About Us

Connect With Us

Resources

Webinar:

How to get the most out of Data and AI
with Ravi Loganathan from Sardine
and President of Sonar

Wednesday, July 24, 2024 at 11 AM ET / 8 AM PT

In this webinar we will cover:

REGISTER NOW!

This webinar is brought to you by:

SardineBanking Exchange