Ahead of the Florida International Bankers Association’s (FIBA) AML Compliance Conference next week in Miami, Basis Technology decided to interview David Schwartz, President & CEO of the association.

From the internet to smartphones, David has seen the impact numerous technological innovations have made on banking during his 30-plus years in the industry. Given his depth of experience, we were interested to hear how he sees AI making its way into this space—and he did not disappoint.

---

How has AI impacted banking so far?

AI has only begun to impact banking, and, as is the case in the early days of any technology, the conversation is filled with buzzwords. “Crypto,” “blockchain,” “machine learning” (ML), and “AI”—there’s a fog of terminology to sort through.

Right now banks are exploring using ML and AI in a plethora of use cases, and they’re doing this to answer a very simple set of questions:

• What are these technologies?
• How do they work?
• And how can they help me?

Has the technology’s impact in the space been overhyped or underhyped?

Definitely not underhyped.

Commercial dialogue around innovations always has a degree of hype. Technology companies are companies after all; they want to sell their product, and they want to make money.

But I wouldn’t call the technology overhyped.

AI isn’t a pet rock.

We are talking about an innovative technology that may or may not be able to solve very important problems. And no one is buying it for no reason

Banks aren’t rushing in, buying, and implementing an AI application just because it’s AI, and they must have it. The industry is being thoughtful. It’s carefully analyzing the benefits of this technology.

If they find it to be useful, efficient, and cost-effective, they’ll implement it.

What does the future hold for this technology and anti-money laundering (AML)?

Of any application area, compliance seems to be drawing the most attention…particularly in customer activity monitoring.

Banks are drowning in false positives. It’s one of the biggest issues in the industry. Current systems are all rules based. The problem is the volume of rules and the number of alerts generated requiring human intervention. Compliance departments are rifling through a mountain of alerts and asking themselves, “Is this suspicious? Do I need to file a suspicious activity report (SAR)?” and it’s taking too long (and too many resources) to answer these questions.

If AI can help reduce this workload, significantly cutting down on false positives (not to mention false negatives), then it will be very useful for banks.  

When financial institutions (FIs) onboard a customer, there’s a standard set of data that has to be collected and checked in order for institutions to fulfill their know your customer (KYC) obligations.

But very soon, I believe AI will enable banks to leverage far more than just name and address. AI-driven systems will ingest web data—from social to news media—and use it to build sophisticated customer profiles that update in real time.

This kind of profile would make it possible for banks to actually know their customers and develop an individualized risk profile.

This profile would be an invaluable resource for transaction monitoring systems. Instead of identifying suspicious activity on the basis of behavioral stereotypes encoded as rules, these systems would be able to spot anomalies based on the activities that are usual or unusual for a particular customer.

How will banks overcome their legacy IT infrastructure to integrate modern technology?

That’s the challenge, isn’t it?

I think banks will soon recognize the necessity of updates and truly begin integrating these innovative technologies.

At the end of the day, it’s about the bottom line. If FIs see benefits superseding the costs, they’ll bite the bullet and update. Unfortunately, in some of the places where AI can do the most good, like compliance, it can be difficult to see the tangible impact on the bottom line.

Compliance departments are seen as cost centers, but they help stop bad actors from laundering money. They save banks from regulatory and reputational risk. While the benefits they generate are more cloudy and diffuse than other business activities, in the long run, the impact is substantial.

If you save yourself from chasing your tail on false positives, you can move those resources elsewhere. If you miss fewer false negatives, you can save yourself millions (or billions) in fines.

If AI can bring radical efficiency and effectiveness to these process, FIs will get the memo and integrate.

One more thing: The compliance professional of the future is not going to be sitting around reading documents. They are going to be tech-savvy. Could even be coders.

In not-too-long, you’re going to see a huge shift in the skills that are needed for this profession.

Besides the enormous effort involved in updating legacy infrastructure, what do you see as the biggest obstacle to AI integration in banking?

Hate to sound like a broken record, but costs are front and center.

Innovation isn’t just necessary for the major players—the big four. It’s also needed at the grassroots level, in small and medium-sized community banks.

These institutions are facing the same issues with compliance staffing and resources that the big four are, but have a fraction of the capital…even relatively speaking. Many of these banks can’t be involved in certain kinds of business because the compliance costs are too high.

If AI is the solution to the AML problem, then they’ll be able to make that investment and upgrade. Or they’ll have to exit higher risk business lines.

Regulatory buy-in is also a potential barrier, but we’re seeing positive developments on that end. The statement issued by FinCEN back at the beginning of December was clearly aiming at encouraging banks to innovate.  

FinCEN seems willing to offer grace to FIs during this transition period, indicating that they won’t necessarily take action if the new applications begin finding suspicious activity the old system would pass on.

Some people are looking at this as an unequivocal green light…as if the regulators are fully on board.

But I’m a bit more cautious.

There’s still a lot to navigate when it comes to making these new technologies compliant, the “black box” problem being a standout, and I think the industry needs to tread carefully as they enter this new era.

How do you see AI changing compliance?

In two respects.

One. Banks are always afraid of missing something, that there was something that ran through their system that they should have caught. Being able to better analyze the data the monitoring systems receive will allow them to  filter out more false positives, while at the same time, because of the enhanced analysis, reduce false negatives.

Two. While false positive reduction means cost savings, false negative reduction means fewer fines plus an added bonus: FIs will be providing better and more actionable information to law enforcement.

It’s a definition case of a win-win.

---

About David Schwartz

David Schwartz is President and CEO of the Florida International Bankers Association (FIBA), a nonprofit trade association dedicated to fostering the growth of international banking through education, advocacy and networking. Its membership includes some 70 financial institutions from 18 countries spanning 4 continents, including the largest banks in Europe, the U.S., Latin America and the Caribbean that are active in international trade and finance.

Prior to joining FIBA, David held executive positions at various financial institutions in the U.S., Europe and Latin America in the areas of Risk Management and Compliance, Wealth Management, Correspondent Banking, Credit and Treasury. Additionally, David is a member of the Board of Managers of QuantaVerse, a leader of artificial intelligence (AI) and machine learning solutions purpose-built for identifying financial crimes.

David received his Juris Doctorate from New York Law School and also has a Bachelor of Arts degree from Florida State University in Tallahassee, Florida. He has completed foreign studies and training in Hong Kong and Paris, France and is fluent in Spanish, Portuguese, French and Italian.