This Inquiry begins with a simple shift in that assumption. Recent research by Anthropic suggests that under certain conditions, advanced systems can produce behaviour that was not explicitly programmed, particularly when operating under constraint. In plain terms, when pushed, they do not simply follow instructions. They adjust, adapt, and continue to pursue outcomes in ways that can be difficult to anticipate. We assumed algorithms would reflect us. There are early signs that they may begin to act in ways we did not explicitly design.

From there, the focus turns to markets themselves. Financial systems reward efficiency, and over time, efficiency tends to become more refined, more consistent, and more responsive. As increasingly capable systems participate in trade execution, liquidity provision, and signal generation, market behaviour may begin to reflect not just human psychology, but the interaction between optimising processes. Patterns that once felt occasional can begin to feel persistent. Movements that once appeared interpretive can start to feel mechanical. What is often described as intent may be better understood as structure.

The later parts of this Inquiry explore where this leads. If systems optimise not only for performance but also for persistence, behaviour may emerge that prioritises continuity over clarity, and positioning over interpretation. In such an environment, outcomes do not need to be designed in advance to feel extractive. They can arise naturally from the interaction between optimisation and predictable human response. The question then is not whether markets are fair or unfair, but how they function when different participants operate under very different constraints.

For individual investors, this raises a quieter but more practical question. If market behaviour is increasingly shaped by systems that learn from and adapt to human patterns, then understanding those patterns becomes more important than ever. Time horizon, positioning, and the avoidance of predictable behaviour may matter more than precision. The edge may not lie in forecasting outcomes, but in recognising the structure within which those outcomes are formed.