CFTC chair Mike Selig says AI agents could outgrow human-era trading rules

Key points

• He asked whether an agent's developer must register and who is liable if an agent "goes rogue"
• An eToro survey published in September found 79% of US millennial investors had used or were open to using AI to pick or adjust investments

CFTC chair Mike Selig says autonomous AI agents could soon hold assets and trade on their own, leaving financial rules built around human participants behind. Speaking at the agency's first Frontier Forum in Washington, DC, on Wednesday, he said a fridge that stakes Ether, buys groceries with the proceeds or even launches a memecoin is "no longer science fiction", Cointelegraph reports.

Selig said the CFTC's framework was written for identifiable market participants carrying out specific functions. An agent that spots trades and executes them for a user raises awkward questions: whether its developer must register, who is liable if it misbehaves, and how regulators can tell on whose behalf it trades. He floated a new form of know-your-customer checks, which he called know-your-agent.

Access and new risks

He argued the technology could widen access to risk management once reserved for large institutions, processing information faster and more accurately than people can. But he warned that agents may react unexpectedly to markets, that similar models could produce correlated trading, and that machine-speed deals complicate surveillance.

The remarks land as younger investors take to AI money tools: an eToro survey published in September found 79% of US millennial investors and 77% of Gen Z investors had used or were open to using AI to choose or adjust investments. Exchanges are already building for it, with Coinbase launching Agentic Wallets in February and Binance adding AI Agent Skills in March.

Read the full story on Cointelegraph →