As technology evolves, so too do the methods used by criminals. To keep pace with these increasingly sophisticated threats, law enforcement agencies are turning to artificial intelligence (AI) technologies to bolster their investigative capabilities.
In a recent interview with MeriTalk, Jarod Koopman, the cyber chief of the Internal Revenue Service’s (IRS) law enforcement arm, shared insights into how AI is transforming the way his team tackles cybercrime, digital forensics, and fraud.
Koopman – who currently serves as the executive director of Cyber and Forensic Services at IRS Criminal Investigation (IRS-CI) – has a background in computer information systems and a two-decade career spanning multiple key roles within IRS-CI. He emphasized the importance of staying ahead of criminals who are constantly adapting their methods.
Koopman noted the explosion in the advancement of cybercrime within the financial sector. “It used to be Bitcoin only – maybe darknet marketplaces with the old school Silk Road days – and it morphed into investment fraud with all the ICO [initial coin offering] frauds,” he said, emphasizing that AI tech is helping his team stay agile.
AI as Force Multiplier in Cyber Investigations
One of the most significant ways IRS-CI is using AI is by leveraging machine learning to analyze vast amounts of data. “Think about the amount of data that we have – from our operational [data] mixed with open-source intelligence – if there’s a way that we can leverage AI to look at all of that to find pattern or to highlight things that are really nefarious or criminal in nature, it just makes us much more efficient in our job,” he said.
For example, his team uses AI tools to run blockchain analytics, asking specific, targeted questions that help uncover nefarious activity. “Almost like a chatbot, to be able to ask questions like, ‘Show me all wallet addresses that interacted with this darknet marketplace,’” Koopman explained. “‘Show me all the ones that have had over $100,000 value in that past year’ … It’s getting more efficient at drilling down into the most egregious cases.”
Previously, Koopman said their leads on cases were like “throwing darts.”
“We’d get leads in from U.S. attorneys, we’d get public [leads], we’d get [leads] we were generating ourselves. It’s not really data driven. It was more what’s popping up and highlighting itself to us versus doing some real analytics behind it,” he said.
Koopman emphasized that while AI is an essential tool, human oversight remains critical. “[AI] is used as a marker to pull out the largest, most egregious cases or investigations that we should be focused on, and then from there, we implement a human-in-the-loop aspect where we still have analysts and agents looking at these cases, saying, ‘Yes, this is a good case that we should get involved in.’”
The Role of AI in Scaling Investigations
IRS-CI’s cyber program, which Koopman helped build from the ground up in 2016, has grown significantly in recent years. From just a handful of agents, the program now includes around 230 personnel, focusing on everything from digital forensics to cybercrime investigations. As the scope of cybercrime has expanded, so too has the need for advanced technologies to process the data involved.
“[Previously,] it was more like automating processes throughout government,” Koopman said. “Whether it’s forms … being able to pull data from one and dump it into another one without having to switch platforms.”
“Now we’re moving into efficiency with more direct outputs that are helping us find the results that we’re looking for. And that’s where the AI starts to learn from the typologies of what we’re truly looking for in order to do our job, and then eventually getting to a place where you can have it start … making decisions,” he said.
With this shift to data-driven investigation, IRS-CI has seen a steady increase in its ability to uncover and prosecute criminal activity. He said that, with the help of AI, his team has gone from working a few dozen cases a year to more than 150 annually, with seizures continuing to rise. In the last three years, IRS-CI has seized over $11 billion in cryptocurrency, a feat Koopman attributes not only to the growing value of digital assets but also to the efficiency AI provides in identifying criminal activity.
AI’s Role in Privacy and Data Sharing
While AI is playing a crucial role in accelerating investigations, Koopman also acknowledged the challenges it brings – especially when it comes to privacy. He explained that his group has to ensure they’re using AI in a way that adheres to strict privacy guidelines, and noted that, within the IRS, they must comply with tax data privacy rules that limit how they can share information.
Koopman is particularly excited about privacy-enhancing technologies (PETs) like fully homomorphic encryption (FHE), which enable the secure sharing of data without compromising privacy. “We have a lot of projects currently going that are using privacy-enhancing technologies and partnering with the likes of Duality [Technologies],” he said. “We leverage those type of technologies to find ways that we can advance the government in the work that we’re doing, to then make that speed and volume much more efficient.”
Using PETs, Koopman’s team can now share data securely with other law enforcement agencies both domestically and internationally, while ensuring that critical data – such as tax information or grand jury material – remains protected. Koopman said it’s all about striking the balance between efficiency and privacy.
The Future of AI in Law Enforcement
Looking ahead, Koopman is optimistic about the future of AI in law enforcement. “Probably one to three years, we’re going to see explosion and growth like never before,” he said. “I really see AI advancing and as the capabilities, even with computing gets higher, eventually with quantum computing and things like that coming into play, where the speed and volume at which we can do things is just tremendous.”
Koopman anticipates more integration across sectors, with AI playing a pivotal role in the convergence of gaming, cryptocurrency, and financial transactions.
“You bring all that together and create this environment that is really opening up into endless possibilities for not only criminals but for government,” he concluded. “Government, and specifically law enforcement, needs to look at it that way as being very innovative and creative. And I think the only way we do that is working with some of these partners and private sector, like Duality, to be able to find creative ways to do that.”
