As a quantitative giant, Two Sigma has long relied on algorithms and data infrastructure, and has previously faced internal model manipulation incidents leading to client losses; Citadel and Point72, led by Ken Griffin and Steve Cohen respectively, have historically faced regulatory scrutiny and market manipulation accusations, with their information systems centrally storing trading strategies and client data, making them high-value targets.

In terms of capital pathways, attackers have expanded their target range through low-cost AI voice cloning, growing from targeting 50 entities in the past to thousands, motivated by the desire to obtain trading signals or extort access; hedge funds are mobilizing internal security teams and external intelligence sharing (such as the newly established FINRA center) to respond, shifting resources from traditional IT budgets to real-time voice verification and employee training.

Similar cases include the 2024 incident where employees of a Hong Kong company were tricked into transferring over $25 million using AI deepfake voices, and the wave of attacks reported this year by Google's cybersecurity unit targeting law firms and professional service companies; currently, Wall Street is in a transition phase from passive defense to proactive AI countermeasures.

This fundamentally represents a technological substitution: traditional social engineering is being replaced by AI-driven real-time voice imitation, with the mechanism being that computing power and open-source voice models significantly lower the forgery threshold, shifting attacks from targeted high-cost to large-scale low-cost, forcing capital-intensive financial intermediaries to reprice their internal trust chains.

ABAB News · Cognitive Laws

Voice can be forged, trust is a vulnerability
AI reduces attack costs, amplifying defense premiums
The most valuable data is locked down first