TL;DR
- Michael Selig, Chairman of the CFTC, warned that existing financial regulations do not adequately address AI agents.
- The regulator raised questions about developers’ legal liability, registration requirements, and identifying the users behind automated transactions.
- Selig noted that AI could make trading and risk management more accessible, but also warned about the possibility of simultaneous trading activity and challenges in market surveillance.
Michael Selig, Chairman of the U.S. Commodity Futures Trading Commission (CFTC), called for financial regulations to be adapted to the rise of AI agents capable of managing assets and executing transactions without human intervention.
Speaking at the CFTC’s inaugural Frontier Forum in Washington, Selig explained that existing rules were designed for identifiable traders performing specific financial functions. However, autonomous systems raise questions that the current regulatory framework has yet to address.
The official used an example to illustrate the potential of this technology: a smart refrigerator could purchase groceries using income earned from staking ETH or even launch its own memecoin.
Selig Raises Questions About Liability for Autonomous Agents
One of the main questions is who should be held legally responsible when an AI agent executes trades on behalf of a user. Selig questioned whether developers should be required to register with regulators and who would bear the consequences of an unauthorized transaction.
He also raised the need to identify the people behind these systems. As an alternative to traditional Know Your Customer (KYC) procedures, he mentioned the possibility of introducing a mechanism called Know Your Agent (KYA).
Adoption of these tools is already growing among retail investors. An eToro survey published in September found that 79% of U.S. millennial investors and 77% of Gen Z investors either use or are willing to use artificial intelligence to select or adjust investments.
Selig argued that these systems can process large volumes of information and provide risk management tools that were previously available mainly to large institutions.
The Risks of Automated Trading and Exchange Initiatives
The CFTC Chairman also warned that autonomous agents could react unpredictably to certain market conditions. If numerous systems use similar models, their decisions could lead to correlated trading activity and amplify price fluctuations.
Execution speed presents another challenge for regulators responsible for monitoring trading activity and managing risks.
The Bank of England had already warned in 2025 that the widespread use of similar artificial intelligence models could amplify financial market disruptions.
Crypto exchanges are also developing tools for these systems. Coinbase introduced Agentic Wallets in February, a solution that allows agents to spend and trade within limits set by users.
Meanwhile, Binance launched AI Agent Skills in March, offering tools to access market information and manage buy and sell orders in the spot market.
