Regulation
The SEC's inquiry into agentic trading is a question about who is liable when an AI agent loses a retail investor's money, not just whether the technology works
As brokerages roll out AI agents that place trades without per-trade approval, the SEC's inquiry is testing an older regulatory assumption: that a human is always the last decision-maker before an order hits the market.
Every prior generation of trading automation, from program trading to robo-advisors, kept a human in the approval loop somewhere close to the trade. Agentic trading breaks that assumption deliberately: the entire pitch to retail customers is that they no longer have to approve each order. That is precisely why the SEC's inquiry matters more than the usual new-product scrutiny cycle. It is not asking whether the software works; it is asking who is accountable when it does not, a question that current broker-dealer suitability and best-execution rules were not written to answer.
The liability chain here is genuinely tangled in a way most fintech launches are not. The AI model developer built the decision logic, the brokerage built the account structure and order-routing rails, and the customer granted standing authority to trade on their behalf, usually via a lengthy terms-of-service acceptance rather than an informed, product-specific disclosure. If an agent misreads a volatile session and executes a string of losing trades, none of those three parties individually did anything a regulator could easily call negligent under existing rules, which is exactly the gap the SEC needs to close before agentic trading scales further.
What makes this inquiry different from a typical rulemaking cycle is timing pressure from the product side. Brokerages are not waiting for regulatory clarity to launch; agentic accounts already have real users and real losses accumulating while the SEC works through the question. That sequencing, product first, liability framework second, is the same pattern that produced messy after-the-fact fixes in payment-for-order-flow and gamified trading interfaces, and there is no reason to expect agentic trading to be an exception.
For a retail investor deciding whether to open one of these accounts today, the practical takeaway is that regulatory protection is still catching up to the product. The account terms, not a regulator-mandated standard, currently define how much recourse a customer has if their AI agent trades badly. That is a materially different risk profile from a traditional brokerage account, and it is worth reading closely rather than assuming existing investor protections extend automatically to autonomous execution.
Read the original: KRDO (Stacker) - The SEC's inquiry into agentic trading is a question about who is liable when an AI agent loses a retail investor's money, not just whether the technology works. Commentary is the independent editorial view of Share Trading; the original article is credited to its publisher.