As banks, fintechs and financial institutions grapple with how to implement artificial intelligence (AI), the U.S. Federal Deposit Insurance Corporation (FDIC) is exploring how to shepherd such integrations while also assimilating changes in the agency’s internal tech stack.

Starting June 14, Sultan Meghji, the FDIC’s inaugural chief innovation officer, and the supervisory body’s tech lab, FDiTech, will play host to banks, fintechs and other entities for a week in order to understand their outlook on using AI and machine learning (ML) systems. Created to promote modernization at financial institutions, FDiTech functions as the supervisory body’s hub for innovation and runs pilot programs, tech sprints focused on emerging technology.

The first leg of discussions between FDiTech and relevant stakeholders will focus on how AI and ML can be used in anti-money laundering compliance, cybersecurity, credit underwriting decisions and automation of back-office processes.

“It was well over 100 different entities that reached out to us to participate,” Meghji told Bank Automation News, referring to the “office hours” announced in April and slated to take place in June. He added that the FDIC is taking a broad-stroke approach toward innovation, both internally and in its supervisory role, and that while the body has been “using artificial intelligence for over five years,” for some use cases, it also deploys older tech that it needs to upgrade.

The FDIC is also one among five financial regulatory bodies which issued a Request for Information (RFI) in March, seeking inputs and comments on the use of AI by financial institutions.

Speaking about concerns that bias may seep into AI-driven credit decisioning systems, Meghji said that it is important to pay attention to how the AI system is designed and the kind of data it ingests, adding that it is equally significant to, “recognize that the vast majority of artificial intelligence in the banking system has nothing to do with making credit decisions.”

While a number of banks, including JPMorgan Chase, Wells Fargo, Citi and Bank of America, have noted that they use AI to guard against fraud and assist customers using digital assistants, few have implemented the technology to underwrite credit.

Noting that while AI should not end up baking fresh bias into the system, Meghji said that it is important to look at how to extricate existing biases, many of which can be quite subtle.

For instance, Meghji said, using a prepaid cell number on a bank application can make certain banking products more expensive to access, since the system assumes that if a user cannot get a personal cell phone account they are, “going to be a more expensive customer.”

“When we talk about bias in the banking system seriously, I would have to take a step back from just talking about artificial intelligence,” he added.

As regulators move towards framing rules on usage of AI in different sectors, Meghji said that while several different approaches have emerged, he feels that the right one isn’t on the horizon yet.

“Frankly, given the different regulatory models around the world, one thing that might be really good in a place like a Singapore, or somewhere like that, might not necessarily make the most sense in the United States. There’s going to be some nuances relative to each jurisdiction,” he said.

Meghji acknowledged concerns about splintering global regulations globally, as noted by automation juggernaut UiPath in its recent response to the RFI. Meghji said that it is unlikely that the United States will become an unwelcoming host for AI technology, though. “I would be very surprised if broad-spectrum regulation against artificial intelligence comes down in the United States, unlike other countries.”

Overall, Meghji said he remains quite bullish on how AI use in financial services could be a “fundamental competitive differential,” in the United States and said that banks exploring how to use the emerging technology should keep a keen eye on vendor due diligence, cybersecurity, and the possible return on investment.

Bank Automation News will host a webinar on automation technology for better risk management and security on Tuesday, June 15, at 11:30 a.m. ET. Register here.