September 17, 2025
Welcome back to Barefoot Innovation, where we search for ways to make the financial system work better through new technology and by rethinking regulation. Today’s show ventures into the world of national security and illicit finance with a leader who doesn’t just want to chip away at money laundering – he wants to make it disappear.
Our guest is David Rushing Dewhurst, a statistician‑economist who has joined the Defense Advanced Research Projects Agency – DARPA – as program manager for a new initiative called A3ML, which stands for Anticipatory and Adaptive Anti‑Money Laundering. A3ML’s mission is audacious: they are setting out to raise the cost of money laundering so high that it becomes unsustainable. Dave argues that demand for laundered money will only dry up when the cost of “cleaning” it equals the value of the funds. If you can raise that cost enough, money laundering will go away.
And he argues that it has to go away. While the financial world thinks of money laundering as a compliance problem, DARPA sees it as something else – a national threat. The scale of the challenge has grown so dramatically that DARPA has elevated it to be a national security priority for the United States. Dave talks about why – from North Korea funding more than half its nuclear program through laundered cryptocurrency, to the role of Chinese laundering networks in the fentanyl crisis. By some estimates, global money laundering is now five percent of the world’s GDP. It is trillions of dollars a year, and is growing fast with, among other things, generative AI making financial crime easier to commit, every day.
In our conversation, we talk about the many things that are not working today. One is the patchwork of suspicious‑activity reporting and siloed fraud teams. Another is that compliance functions as a cost centre that rewards box‑checking and over‑reporting, which in turn is overwhelming government agencies with low-value information while missing the actual crime.
The solution to these kinds of technology-driven problems is technology. The forces of good, including law enforcement and the financial industry, need the technological (and legal) tools to enable them to share data on a huge scale, in order to detect the activity patterns that signal crime and money laundering. And in order to do that safely, they need ways to keep all this information protected and secure. This means, as Dave points out, that we should not centralize it in a big honey pot where the data can get stolen or abused. We need new methods.
That is what A3ML is trying to create, using DARPA’s unique system in which people like Dave come into the government and run prize competitions that challenge innovators to take on very hard problems. DARPA is developing high‑speed graph algorithms that identify suspicious patterns across distributed data stores, without sharing sensitive information.
Think about it. If A3ML succeeds, it will mean that, a few years from now, we will have a revolutionary new way to fight money laundering – even, again, to make it so expensive that it’s not worth doing.
If that solution exists, how will it get implemented? As Dave explains, DARPA doesn’t implement anything. Instead, the action agenda will fall to the kinds of people who listen to this show. Our whole edifice of anti-money laundering law and compliance activity will need to be rethought. That will, in my view, have to include modernizing the whole system we use for managing identity including, again, protecting privacy. (The digital identity need is also surging due to the advent of financial AI agents – check out my recent episode with Sean Neville on that!). It seems like now is the time to think about what reforms may be needed and who will champion them, balancing security with civil liberties.
Three more quick thoughts before we hear from Dave. One of the things I love about this conversation is the focus on human incentives and motivation, rather than process. So many of our regulatory strategies totally ignore consideration of incentives and instead just assume that if we require people to follow a certain process, our regulatory goal will be met. In many areas, that was probably the best approach available when a law or rule was originally adopted. Today, very often, it isn’t, because today, we can make complex financial systems much more transparent than was possible in the past. Opaque systems give people incentives to do things they want to keep hidden. Transparent systems penalize that. As new technology enables more and better use of digitized data, we are going to find we can make big progress on a whole range of regulatory goals, using a version of Dave’s logic – that it will become difficult and expensive to conceal bad behavior.
The second thing I love is that this is a rare conversation in the financial regulatory world, because it aims to actually solve the problem. Not reduce it, not push harder against it with the same old tools, but actually to solve it. Here again, technology is making it possible to imagine big leaps forward on a wide range of financial regulatory goals. Often, the secret to doing so is, again, to leverage enhanced transparency. We should be thinking boldly, and freshly, about all our financial regulatory strategies. That is what we do at AIR.
Last but not least, this episode is a great example of our mantra, at AIR, that the answers to the complex problems we grapple with almost always require breaking the walls of silos. The whole financial and financial regulatory ecosystem is massively siloed. No one in it has the ability to solve any of the big challenges from inside their own walls. If DARPA finds a way to fix the money laundering problem, it will be because it’s coming at it from a new perspective, using new tools that are not in the toolbox of the traditional players. We all need to be searching for answers in new places today, as technology changes our world so fast. That means we need to be talking with different people.
As always, I’d love to hear your thoughts. Could this approach end money laundering as we know it? What other regulatory problems might be ripe for a similar rethink? Join the conversation.
David Rushing Dewhurst, Ph.D., joined DARPA as a program manager in April 2024 to design, execute, and transition programs at the nexus of technology and economic strategy.
His research interests include financial intelligence, the design of economic mechanisms and financial assets, capital markets, and payment systems. Before joining DARPA as a program manager, Dewhurst served as a technical advisor for government research and development programs at the agency and was a Capstone Fellow at Yale University. Dewhurst also worked as a senior scientist at a defense research and development firm, a risk management data scientist in the financial industry, and a computer scientist at a federally funded research and development center.
Upcoming Podcasts
We have some wonderful shows coming up, including one with Saket Narayan of AWS about how to modernize the regulators’ own technology, and a show with Carole House and Amanda Wick on financial crime, and one with Fourish Ventures’ Tilman Erhbeck on (wait for it)… how to change the world.
AIR Events
Calling all Regulators / Meet in London: We are excited to be offering a four day learning program in London in October, inviting regulators and central banks from around the world. The program shares AIR’s Innovation Elements methodology that regulators can use to build your innovation capabilities. We have room for a few more regulators to participate, and we’re also seeking sponsors to support their participation, so reach out if you’re interested.
Stablecoin Roundtable: On September 16 in Washington, D.C., we’re hosting an exclusive roundtable on stablecoins. Attendance will be limited but if you’re interested in receiving an invitation, let us know.
We have also posted a video from our June Battlefront exercise on AI and Fraud. Please check it out.
Upcoming Speaking Events by AIR’s team
I’m looking forward to speaking at Money 2020 in the fall in Las Vegas, doing a fireside chat with Rodney Hood, former Chairman of the National Credit Union Administration and former Acting Comptroller of the Currency. My colleague Shelley Anderson will speak later this month in Johannesburg at the Global SME Finance Forum 2025 — focusing on inclusive growth and SME access to finance. I’m also looking forward to speaking at an event that the Consumer Bankers Association is holding on agentic AI.
Here’s a full list of speaking engagements and events.
And remember, please give us a 5 star rating on your favorite podcast platform.
Stay informed by joining our mailing list
