Congress Debates Bank Secrecy Act Overhaul as AI Fraud Surges 500% and Crypto Demands Reform

Crypto, Banks, Policy Experts Press Congress to Modernize Bank Secrecy Act - Decrypt

Witnesses were split between full repeal, targeted reform, and modernization with stronger information-sharing. Crypto executives, policy researchers, and national security experts testified before a House subcommittee on Thursday on how to modernize anti-money laundering laws for an era of AI and digital assets. The House Financial Services Committee's National Security, Illicit Finance, and International Financial Institutions Subcommittee held a hearing on Modernizing the BSA for Financial Crime in the 21st Century, revisiting the Bank Secrecy Act, the 1970 law that requires banks and financial institutions to report suspicious activity and large transactions. The hearing landed as crypto firms, banks, and civil liberties groups push to refocus the BSA on actionable intelligence over reporting volume, while the Trump administration broadens its reach over non-citizen customers. Redbord warned that AI-enabled scam activity surged 500% over the past year while illicit funds now move across wallets within 24 to 48 hours, compressing response windows to the point where "retrospective reporting frameworks are structurally incapable of generating a response in time" because "the framework that helped us win yesterday will not be enough to win today." The Bank Secrecy Act is the backbone of U.S. anti-money laundering law, requiring banks and crypto firms registered as money services businesses to file suspicious activity reports, currency transaction reports for amounts over $10,000, and to verify customer identities. Subcommittee Chairman Warren Davidson (R-OH) opened by calling the BSA a "bloated surveillance machine demanding endless reports without delivering proportional results," noting institutions file nearly 5 million SARs and 21 million CTRs annually. Also, he noted how institutions should hold "the least amount of information that they need on an individual customer in order to make a decision" about illicit risk, warning that every new database is "a honeypot" for ransomware groups and state hackers. During the hearing, Cato Institute researcher Nicholas Anthony said the problem with BSA was not inefficiency but surveillance itself, saying "the history of financial surveillance has been a history of ever-moving goalposts" from tax enforcement to "fraud and immigration," while laying out options ranging from inflation-adjusting thresholds to repealing the BSA regime entirely. John Court, general counsel at the Bank Policy Institute, backed reform rather than repeal, calling Treasury's proposed AML rule "a vast improvement" and urging higher reporting thresholds, simpler filings, risk-based oversight, and explicit approval for banks to use AI in transaction monitoring. Atlantic Council Senior Fellow Carole House pushed back on deep cuts to the framework, noting that reducing compliance burden should not come at the cost of "opening the door to adversaries seeking to harm Americans and U.S. national security interests." AI was a point of consensus among most witnesses. Davidson, Redbord, Carole House, and Court all backed broader use of machine learning and AI in transaction monitoring. Redbord pushed hardest, telling lawmakers that "AI investigative tools can compress weeks of manual analysis into minutes" and calling for federal funding of AI-native investigative tools across IRS-CI, FinCEN, OFAC, FBI, DEA, Secret Service, and HSI. The hearing arrived only days after President Donald Trump signed an executive order directing regulators to strengthen customer due diligence and customer identification requirements under the BSA while expanding scrutiny around account ownership and financial risks associated with immigration enforcement. The order directs Treasury to tighten BSA customer due diligence rules and flag risks tied to ITIN use, off-the-books wages, and foreign consular IDs, while also asking the CFPB to weigh potential deportation risks in lending decisions.

House Hearing Puts AI Fraud at Center of AML Overhaul | PYMNTS.com

By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions. At issue was whether the Bank Secrecy Act should evolve into a faster, more targeted and technology-enabled system, or whether reducing reporting obligations risks creating new blind spots just as artificial intelligence (AI), synthetic identities and crypto infrastructure reshape financial crime. At the House Financial Services Subcommittee hearing, "Modernizing the BSA for Financial Crime in the 21st Century," lawmakers and witnesses broadly agreed on one point: the threat environment has changed. The disagreement centered on how much of the existing framework should change with it. Witnesses included John Court, executive vice president, general counsel and chief operating officer at Bank Policy Institute; Ari Redbord, global head of policy at TRM Labs; Nicholas Anthony, research fellow at the Cato Institute's Center for Monetary and Financial Alternatives; and Carole House, senior fellow at the Atlantic Council and former White House and Treasury official. Opening statements framed the competing philosophies. Subcommittee Chairman Warren Davidson, R-Ohio, argued that the current framework produces enormous reporting volumes without equivalent investigative value, pointing to millions of Suspicious Activity Reports (SARs) and Currency Transaction Reports (CTRs) generated annually and questioning whether thresholds created decades ago still reflect modern financial activity. He said that regulators should move away from what he described as defensive compliance and toward actionable intelligence supported by AI tools. Ranking Member Joyce Beatty offered a different warning. She stated that financial crime remains deeply embedded across the system and said modernization should not become a vehicle for weakening anti-money laundering controls at a moment when fraudsters increasingly use crypto and AI-enabled techniques. That tension carried over into witness testimony. Court argued that Congress already attempted to reset the framework through the Anti-Money Laundering Act of 2020, but implementation has lagged while supervisory expectations remained focused on process and documentation. For years, he testified, banks have faced "an incessant focus on technical compliance and examiner benchmarking" rather than flexibility to use risk-based judgment to identify serious criminal activity. During questioning from Rep. Frank Lucas, Court said threshold reform would help but would not solve the deeper problem. "Modernizing the threshold would certainly help," Court said. "Currently we have thresholds that are way too low that are creating just too much volume." But he argued that the larger opportunity is to change incentives so institutions focus on higher-value intelligence rather than examiner-driven process requirements. Court also urged FinCEN to take a more active oversight role in examinations and called for modernization of SAR and CTR rules to simplify reporting and align expectations with actual risk outcomes. If reporting volume represented one side of the hearing, AI represented another. Redbord argued repeatedly that financial crime now moves too quickly for systems designed around retrospective reporting. In written testimony, Redbord warned that fraudsters are using AI to generate synthetic identities, automate fraud and move value across digital asset networks at speeds traditional investigative processes cannot match. He cited North Korean digital asset theft and industrialized scam operations as examples of how technology has changed the economics of financial crime. His argument to lawmakers was not to slow innovation but to match it. During questioning, Lucas asked Redbord whether Treasury and law enforcement agencies would benefit from collecting less low-risk information. Redbord answered yes, but only if institutions become better at producing intelligence. "We'd have fewer SARs, but they would be more targeted, more focused," he said. He argued that traditional SAR narratives often isolate individual transactions when modern investigative tools can map entire criminal networks. "What we could do today is build out networks with SARs. We could really focus on where the true risks are to the financial system and to that financial institution." That position aligns with Redbord's written testimony that included a proposal for modernized reporting structures that move from transaction-by-transaction alerts toward network-level intelligence products that law enforcement can operationalize more quickly. House observed that that modern financial crime deliberately fragments activity across institutions and jurisdictions, making cross-institution visibility and structured information sharing increasingly important. She proposed more machine-readable financial crime data standards and expanded information-sharing mechanisms. Identity emerged as one of the hearing's clearest themes because it sat at the intersection of AI, fraud prevention and privacy. Witnesses told the lawmakers that today's compliance burden reflects broken identity infrastructure rather than excessive regulation. Financial institutions repeatedly verify information government systems already possess, often through expensive and increasingly unreliable document-based processes. House warned that deepfakes and generative AI are creating what she described as an authenticity crisis for traditional KYC systems and pointed to FinCEN analysis identifying $212 billion in identity-compromise-enabled suspicious activity. That theme surfaced directly in questioning. Chairman Davidson asked House whether she supports digital identity. House answered: "I supported digital identity, privacy preserving and dignity preserving. Yes." Davidson then turned to Redbord and asked how modernization could protect privacy without turning financial records into broad surveillance infrastructure. Redbord argued that modernization should mean collecting less information but using it more effectively. "What we need to do is ensure that lawful actors are able to transact in a secure and private manner," he said. Financial institutions should have "the least amount of information that they need on an individual customer in order to make a decision" about illicit risk. That answer reflected one of the hearing's more notable areas of overlap: despite disagreements over reporting thresholds and enforcement, multiple witnesses argued that modernization should improve intelligence quality without automatically expanding data collection. The hearing ultimately left unresolved where Congress draws the line. But the testimony suggested that the next phase of AML policy may be less about generating more reports and more about deciding which intelligence actually changes outcomes before illicit funds disappear.