Law360 (September 25, 2026, 11:46 AM EDT) -- Artificial intelligence is playing an increasingly important role in policing prediction markets, and attorneys say that regulators and investigators are starting to effectively use the same technology that would-be bad actors have leveraged to try to score big with their bets.

Federal prosecutors have launched several high-profile enforcement actions this year involving predictions allegedly made with insider information. The first-ever criminal insider trading case involving a prediction market came in April, when the U.S. Department of Justice and the U.S. Commodity Futures Trading Commission alleged that an active-duty U.S. Army service member used classified information to predict the capture of Venezuelan President Nicolás Maduro on Polymarket.

A month later, a grand jury indicted a Google software engineer for allegedly misappropriating confidential, proprietary Google search-trend data to net $1.2 million betting on Polymarket's "Year in Search" contracts.

Kevin Frankel, head of the state attorneys general practice at Benesch Friedlander Coplan & Aronoff LLP in San Francisco, said that the government may have been slower out of the gate to use AI to target prediction market insider trading, but the gap seems to be closing.

"They can use the same sort of AI tools that folks are using to beat the markets to find the people beating the markets," Frankel said.

Much of the investigative work involves spotting suspicious trading activity, odd patterns and other anomalies by using AI to scan a huge number of trades across various platforms.

Frankel noted that CFTC director David Miller has said that policing illegal activity on prediction markets is among the commission's top priorities, setting up a protracted battle with cutting-edge technology playing a starring role.

"Big picture," he said, "both sides of the equation, the traders and the regulators, are pulling out their AI arsenal."

Adria Perez, a white collar partner at Reed Smith LLP in Atlanta who represents companies, said that her clients often receive letters from the DOJ flagging a potential violation of the law after finding an anomaly or inconsistency that the government feels warrants a closer look, including activity related to trading on prediction markets.

"It's sort of forcing our clients to fight AI with AI," Perez said. "[Companies] are using AI internally to try to understand these inconsistencies and anomalies first. We are talking about AI every day, all day, whether it's with the government or it's with our clients."

Philadelphia-based Duane Morris LLP partner AJ Rudowitz likened the technology to the newest version of the old-school casino pit boss who scanned the floor to make sure everything was in order. Eventually, casinos began to use high-tech cameras and algorithms to spot cheating.

"With AI, it's getting more and more real-time, and more and more accurate," Rudowitz said. "You're not seeing as many false positive hits, and you're able to get to it quicker and see a broader picture."

Reed Smith's Perez said one of the challenges her clients are facing is that, while AI can flag patterns, it cannot always fully capture the circumstances surrounding certain activity.

"We are constantly trying to provide context," Perez said. "We are trying to explain 'yeah, it looks like an anomaly on the black-and-white, but let me explain to you why and how that happened.'"

Perez said investigations are getting triggered more quickly when enforcement agents use AI, and those "critical conversations" between attorneys and the government are happening earlier in cases. Things could proceed even faster as regulatory agents and companies both get better at using AI tools.

"From the prediction market standpoint, they are going to get better and better at detecting patterns and being able to stop and identify users, but bad actors are learning too," Perez said. "Whether it's using a different VPN or thinking about ways to do smaller bets, everyone is going to get smarter."

Prediction markets have been self-policing in high-profile instances. Kalshi issued lifetime bans to several individuals, including former U.S. Rep. George Santos for making a bet that he would not show up at the State of the Union address — an outcome of which he was in complete control.

Rudowitz said that AI is helping platforms spot red flags so fast that, in some instances, they prevent the bet from taking place at all. He still expects this to be a busy area for lawyers moving forward.

"There's a lot of different potential issues, and attorneys are going to be needed to make sure the regulatory schemes are being adhered to," Rudowitz said. "You're also going to need attorneys who are really aware of these laws to help any athletes who might get caught up in allegations."

While instances of athletes being tied to prediction markets are rare, in August it was revealed that the NCAA disciplined two Mississippi State football players who placed Super Bowl bets on Kalshi in violation of NCAA rules. The wagers were flagged by ProhiBet, a tool the NCAA uses to detect possible betting activity.

Benesch's Frankel said that the insider traders have the edge right now when it comes to using AI to beat the markets, but said "the regulators do eventually catch up."

"The statute of limitations is generally five years, not five minutes," he said. "We saw this with crypto. There was a time when everyone thought 'this is better than cash' for illicit activity. Within a couple of years, it flipped to the government saying 'we would rather criminals use crypto than cash, because they leave this blockchain, and our guys can crack it.''

--Editing by Bruce Goldman.

For a reprint of this article, please contact reprints@law360.com.