DARPA wants AI to help anticipate money laundering activities before they happen

Will DARPA’s precrime algorithms remain in the hands of the Pentagon? perspective

DARPA wants AI to help anticipate future money laundering and other illicit financial activities before they even happen.

With the proposed Anticipatory and Adaptive Anti-Money Laundering (A3ML) research program, the US Defense Advanced Research Projects Agency (DARPA) is looking to eliminate global money laundering entirely, according to the A3ML special notice.

“DARPA seeks to revolutionize the practice of anti-money laundering through its A3ML program,” the program description reads.

“A3ML aims to develop algorithms to sift through financial transactions graphs for suspicious patterns, learn new patterns to anticipate future activities, and develop techniques to represent patterns of illicit financial behavior in a concise, machine-readable format that is also easily understood by human analysts.

“The program’s success hinges on algorithms’ ability to learn a precise representation of how bad actors move money around the world without sharing sensitive data.”

The A3ML special notice highlights a federal indictment alleging that Chinese underground banking launders money for the Sinaloa cartel and that half of North Korea’s nuclear program is financed by laundered funds.

DARPA makes no mention of digital currencies or cryptocurrencies; however, the money laundering sources cited in the special notice do.

For example, the document cites a Congressional Research Service report that has a section dedicated to the question of “Cryptocurrency Regulation,” which states:

“Whether (or to what extent) the digital asset industry requires enhanced AML regulation and how certain financial technology (fintech) falls within the scope of US sanctions were issues contemplated by the 118th Congress and may be further addressed in the 119th Congress […]

“The [Congressional] bills also raised complex policy questions regarding the desired scope of AML regulations for virtual asset service providers, anonymizing services, virtual currency or digital asset kiosks, and other decentralized service providers, such as un-hosted wallet providers, digital asset mixers, miners, validators, and other nodes in the cryptocurrency-related ecosystem.”

Leading the proposed A3ML program is David Dewhurst, a fan of beginning sentences with lowercase letters, who joined DARPA as a program manager in April 2024 but worked on DARPA projects in various roles during his almost three-year tenure at Charles River Analytics, according to his LinkedIn profile.

In a recent LinkedIn post, Dewhurst mentioned that the goal was “to stop adversaries from laundering money to evade sanctions, buy weapons, and fund drugs that kill Americans,” and that the technical hypothesis for A3ML was that “illicit finance tactics, techniques, and procedures can be algorithmically extracted from diverse data sources and represented in a generic sharable form.”

Do these illicit finance tactics, techniques, and procedures include the use digital assets and cryptocurrencies?

If and when the A3ML program is launched and is successful, will the algorithms remain in the hands of the Pentagon, or will they also be shared with allies or become commercially available?

Apart from the immense good that can come from eliminating money laundering that funds the killing of Americans, could the tools and tactics developed contribute to creating precrime algorithms that could one day be weaponized against American citizens?

What could this mean for law-abiding crypto enthusiasts and/or Bitcoin holders in the future?

DARPA intends to hold an industry day for the A3ML program in the coming weeks.

Image: Image by macrovector_official on Freepik