The head of the U.S. Commodity Futures Trading Commission (CFTC) said financial regulation should be redesigned for an environment in which autonomous artificial intelligence (AI) agents can hold assets and trade directly.

Cointelegraph reported on Oct. 29 that CFTC Commissioner Michael Selig (마이클 셀릭) said at the first Frontier Forum in Washington that existing regulatory systems are built on the assumption of human traders. He said they must now address situations in which AI agents independently carry out financial activity.

He said it has become possible to imagine scenarios in which a refrigerator stakes ether and automatically buys groceries with the proceeds, or issues its own meme coin. He said regulators must answer how they would supervise such activity. “We are gathered to discuss technologies that are rapidly changing financial markets,” he said. “This is a technology as real as the printing press, the steam engine and the internet,” he said.

The core issue is who bears responsibility. Selig mentioned that if an agent slips beyond control, there are questions over who is liable. He also said regulators may need an “agent verification” system, different from existing know-your-customer systems, to identify on whose behalf multiple agents are trading.

Expectations and concerns over the use of AI were also raised. Selig said AI could provide individual investors and small businesses with risk-management capabilities previously available only to large institutions. He warned that as AI becomes more autonomous, unexpected market reactions could emerge. He said trades could become one-sided if agents using similar models make similar decisions. He also noted that ultra-high-speed automated trading could make market surveillance and risk management more difficult.

Similar concerns have come from other institutions and the industry. The Bank of England's Financial Policy Committee warned last year that widespread use of similar models could amplify market shocks. Animoca Brands Chairman Yat Siu also said that if many users ask AI the same trading question, they could all get similar answers.

Crypto trading infrastructure is rapidly adopting autonomous trading tools. Coinbase unveiled an “agentic wallet” in February that allows AI agents to execute spending and trading within user-set limits. Binance also introduced features for AI agents in March and began supporting the provision of market data and spot order placement and management functions.

AI adoption among younger investors is also cited as a factor accelerating regulatory debate. In a survey released by eToro in September, 79 percent of U.S. millennial investors and 77 percent of Gen Z investors said they already use AI or are willing to use it to choose investments or adjust portfolios. As a result, regulatory standards to be applied across automated financial services are increasingly likely to directly affect protection for individual investors and market stability.

In this context, the CFTC's concerns are not limited to the crypto market. As asset holding, order execution and risk management shift to AI agents, regulators face the task of redefining registration requirements, assignment of responsibility, monitoring methods and verification procedures. As AI moves beyond a tool assisting financial participants and becomes closer to an independent actor, related regulatory discussions are expected to gather pace.