Predictive Policing and the Minority Report Paradox
Forecasts of risk can shape where enforcement happens and what data is collected next.
The Mechanism: Predictive-policing systems analyze historical records or incident data to identify places or people considered more likely to experience or be involved in crime. Their outputs depend on recorded activity, definitions, and the decisions made by agencies using them.
The Philosophical Pivot: A forecast is not neutral when it changes where police look. Increased scrutiny can produce more recorded incidents, which may then appear to validate the original prediction while obscuring unobserved harm elsewhere.
The Horizon: Public safety cannot be reduced to the efficient allocation of surveillance. Any use of prediction must be judged by transparency, disparate impact, community consent, and whether it improves safety without turning probability into suspicion.
A person should not have to disprove a statistical risk assigned to a neighborhood or demographic group. Fair systems require a path to challenge how data was gathered and how the output is used.
The promise of prevention becomes unjust when the forecast itself becomes grounds for treating someone as guilty.
When a forecast directs the search, it may begin to create the evidence it claims merely to find.
If I am labeled a risk by data gathered under unequal scrutiny, what would make that judgment fair or contestable?