FinCEN’s new AML rule rewards banks that collaborate

August 18, 2026

FinCEN and the federal banking agencies have proposed a rule that could reshape how banks think about experimenting with new anti-money laundering technology, and the message buried inside it is hard to miss once you look for it: collaboration will count in a bank’s favour.

According to Consilient, the notice of proposed rulemaking (NPRM), issued on 7 April 2026, reforms the Bank Secrecy Act’s AML/CFT programme requirements.

Consilient recently discussed FinCEN’s 2026 NPRM, and what it labels the AML collaboration imperative.

It stops short of mandating any specific technology, but it names categories of innovation, including federated learning, artificial intelligence, and other advanced monitoring tools, and states that experimenting with them will not expose a bank to enforcement risk. The proposal also builds on the AML Act of 2020’s stated purpose to “encourage technological innovation and the adoption of new technology by financial institutions to more effectively counter money laundering.”

Under the proposed rule, a federal banking regulator would need to give FinCEN’s Director at least thirty days’ written notice before taking a significant AML/CFT supervisory action against a bank. In weighing whether to intervene, that director must consider whether the bank is using innovative tools capable of producing demonstrable evidence of its programme’s effectiveness.

Three practices map onto that encouragement. Federated learning lets models train across multiple banks’ data without records leaving individual institutions, addressing the fact that no single bank sees enough of any one typology to build a fully representative model. Synthetic data, generated from patterns in filed SARs, supplements thin real-world samples of rare typologies without exposing customer data.

A comparative dashboard benchmarks a bank’s own filing patterns against peer institutions and FinCEN’s own database, giving compliance officers, and examiners, an objective measure rather than a narrative self-assessment.

These approaches already have a track record elsewhere. In healthcare, the Federated Tumor Segmentation initiative trains diagnostic models across more than 30 hospitals without sharing patient records. In finance, the BIS Innovation Hub’s Project Aurora found collaborative analysis detected up to three times more complex money-laundering schemes and cut false positives by as much as 80% versus siloed monitoring.

Consilient, whose platform combines federated learning, synthetic data and cross-institution benchmarking, argues these tools can be adopted gradually rather than as a wholesale system replacement. The comment period on the NPRM (Docket FINCEN-2026-0034) closed 9 June 2026, with a proposed twelve-month implementation window following any final rule.