New in Law360: Insider Trading Doctrine in an AI Market
I recently published an article in Law360 (linked below) that uses the SEC’s settlement with Virtu as a jumping-off point to think through a question the doctrine has not fully confronted yet.
Virtu itself is not an AI case. At its core, it is a fairly traditional controls matter involving access to customer order flow. What makes it interesting is what happens if you take the same underlying facts and change just one variable. Instead of the data sitting in a database or being accessed by humans, imagine it being used to train an AI model that informs trading decisions or, in more advanced settings, trades on its own.
That thought experiment raises important questions. If a model is trained on material nonpublic information, does the model itself become a vessel of MNPI. What does it mean to trade on the basis of MNPI when the decision-maker is a system rather than a person. And how do existing insider trading theories, built around human intent and traceable information flows, map onto that reality.
The article is less about Virtu as a case study and more about using a familiar enforcement posture to explore how those same principles may apply as AI becomes embedded in trading, surveillance, and compliance functions.
Really interesting issues at the intersection of enforcement, market structure, and emerging technology.
You can read the full Law360 article here:
