CFTC Chair Selig raises questions about AI agents in crypto markets

What happened

CFTC Chairman Michael Selig said autonomous AI agents raise novel regulatory questions for cryptocurrency markets at the commission's first Frontier Forum in Washington, DC on Wednesday. Selig described scenarios where AI systems autonomously manage funds, execute trades, and even launch new tokens on blockchain networks. He referenced hypothetical examples such as connected devices earning yield through cryptocurrency staking or creating new digital assets, illustrating how the technology blurs traditional boundaries between users, developers, and market participants.

Selig noted that major AI systems including Claude, ChatGPT and Gemini can already interface with other software to analyze market data, and some brokerages have begun integrating these tools for customer use. He pointed to the emerging practice of users granting AI agents direct access to bank and brokerage accounts to manage investments. On blockchain networks specifically, agents can now hold and trade cryptocurrency without intermediaries like banks or brokers, a capability that expands the scope of autonomous financial activity.

The CFTC chairman acknowledged investor appetite for AI-assisted trading, citing an eToro survey finding that 79 percent of US millennial investors and 77 percent of Gen Z investors used or were open to using AI for investment decisions. Selig emphasized that his remarks represented his personal views and did not necessarily reflect the full Commission's position, according to Blockonomi.

Why it matters

The regulatory framework Selig described was built for "identifiable market participants," but autonomous software operates outside that model. Selig posed critical questions about how existing rules apply: Does creating an AI agent trigger registration requirements? Who bears responsibility if an agent executes trades that damage markets or harm customers? Should regulators implement "know-your-agent" verification processes similar to customer identification rules? These questions lack clear answers under current law. Selig also raised concerns about systemic risk. Agents built on similar models could generate comparable trading signals and execute them at speeds that complicate market surveillance and potentially amplify price movements during volatile conditions. The Bank of England's Financial Policy Committee previously warned that widespread reliance on similar AI models could amplify market shocks, according to Blockonomi's reporting. Selig acknowledged that existing regulatory proposals like Regulation Automated Trading, adopted after the 2010 Flash Crash, would be ill-suited for today's autonomous trading environment.

Background

The CFTC has been developing a regulatory framework for cryptocurrency markets even as Congress debates comprehensive market-structure legislation. In October, the agency opened a consultation on how crypto platforms could operate under a voluntary federal registration pathway, according to Crypto.news. That process distinguishes between spot trading, which generally remains subject to state money-transmission laws, and leveraged retail trading, which would require CFTC registration. The agency is also considering a new designation for crypto asset markets as a subcategory of designated contract markets. Selig has stated that the commission wants regulations capable of preventing fraudulent schemes before customers suffer losses. The October consultation seeks public input on customer fund protections, token listing standards, conflicts of interest rules and proof-of-reserves requirements.

Cryptocurrency exchanges are already deploying AI agent capabilities. Coinbase launched Agentic Wallets in February 2026, allowing agents to spend and trade cryptocurrency within user-set limits. Binance introduced AI Agent Skills in March 2026 to provide market data and tools for spot trading. Earlier this year, the CFTC solicited public comment on compute derivatives, which could allow firms to hedge the cost of computing power, with the commission receiving more than 100 comments.

What to watch

The CFTC indicated it is actively examining how AI and autonomous systems should be regulated in financial markets. Selig noted that CFTC staff already use AI tools for market surveillance and enforcement. The agency's next steps likely include determining whether existing regulatory categories apply to AI agents or whether new rules are needed. The commission will also need to coordinate with the SEC and other regulators as autonomous systems become more prevalent across financial markets. Congressional action on comprehensive cryptocurrency legislation could also influence how the CFTC approaches AI agent regulation.

Questions readers ask