If there is a lack of collaboration between regulators and the industry to promote a competitive landscape, the surge of artificial intelligence innovation could potentially overwhelm smaller banks and fintech companies.
A notable aspect of DC Fintech Week, which concluded in Washington, D.C. on Wednesday, was the participation of key figures such as Federal Reserve Vice Chair for Supervision Michael Barr, Acting Comptroller of the Currency Michael Hsu, and Superintendent of the New York State Department of Financial Services Adrienne Harris. These panelists engaged in discussions regarding the challenges and opportunities that the government faces in relation to AI.
During the conference, Poorani Jeyasekar, a director at Klaros Group, expressed appreciation for the authorities’ thoughtful consideration of AI technology and their efforts to comprehend its functioning.
Some attendees raised concerns about the challenges associated with hiring and budgeting that businesses, particularly smaller ones venturing into emerging technologies like AI, are encountering.
Elizabeth Kelly, who previously founded the financial planning and expense management company United Income before its acquisition by Capital One Financial and currently serves as the president’s special aide for economic policy at the White House, shared insights from her experience working in a financially constrained finance sector.
In a discussion on conceptual AI versus legacy AI, Kelly highlighted the substantial costs involved in developing large language models, emphasizing the limitations and expenses associated with accessing data, computing power, and semiconductors.
She emphasized the constraints these expenses impose on business models, underscoring the importance of a robust and competitive ecosystem to navigate such challenges effectively.
A recent executive order from the White House has called for stricter control of AI by federal agencies, focusing on aspects like accuracy, intellectual property rights, and bias mitigation.
Claudio Calvino, a senior managing director specializing in criminal and litigation consulting training at FTI Consulting, pointed out the existing disparities in the AI market, particularly in areas such as data accessibility, specialized expertise, and computational resources. He highlighted the growing emphasis on relational AI due to its increasing popularity and media coverage.
Tim Massad, former president of the U.S. Commodity Futures Trading Commission and current research fellow at the Kennedy School of Government at Harvard University, echoed Kelly’s concerns about precision in AI models. He cautioned that businesses may deploy models without fully grasping their implications, potentially favoring established companies with abundant data resources over smaller entities with limited budgets.
Massad also highlighted the challenges faced by authorities in keeping pace with rapidly evolving technologies, suggesting alternative approaches to resource utilization beyond mere access to models and supervised firms’ internal audits and testing records.
One proposed approach involves requiring regulated entities to provide evidence of specific testing or certification by a third party or adherence to designated standards.
Massad suggested the potential role of self-regulating organizations in addressing these issues, drawing parallels to their contributions in establishing frameworks for stocks and derivatives markets.
Adrienne Harris emphasized the delicate balance regulators must strike between overseeing AI systems effectively to provide support without stifling early-stage innovation, a sentiment that resonated with Jeyasekar.
Harris, who has overseen digital currency regulations in her role since 2015, updated regulations earlier this year to enhance standards for valuation, security protocols, and anti-money laundering measures.
She noted the inherent tension between consumer protection, business interests, and fostering a competitive marketplace, highlighting the complexities involved in regulatory decision-making.
Jeyasekar observed a greater level of support and regulatory interest in AI compared to cryptocurrency within the financial sector.
Various participants discussed the practical applications of AI in different contexts.
Steve Holden, senior vice president of single-family analysis and modeling at Fannie Mae, highlighted the potential of AI in structuring unorganized data efficiently, citing an example where Fannie Mae automated loan screening processes using AI to extract relevant information from applicant bank statements.
Ardent Venture Partners, a business-to-business banking investment firm, introduced a model named ArgentGPT, which leverages conceptual AI and large language models to streamline the analysis of numerous businesses and owners.
The New York State Department of Financial Services is developing a chatbot for its website to provide simplified responses to visitor inquiries, aiming to enhance user experience and accessibility to regulatory information through AI technology.
Harris offered practical advice to regulated entities venturing into innovative technologies, emphasizing the importance of transparent communication with regulators to avoid surprises and encouraging a mindset shift towards embracing disruption in the financial sector.
She highlighted the evolving perspectives within the industry, urging organizations not to overlook the insights of experienced professionals transitioning from regulatory roles to private sector positions.
