engleză
Dana Cross built ORCA to remove human bias from corporate layoffs.
The model was supposed to make the difficult decisions fairly-measuring performance, cost, and redundancy without favoritism or emotion. But while preparing a workforce reduction for Kellner Foods, Dana discovers a pattern hidden inside ORCA's results.
Employees with long commutes. Employees who used family or medical leave. Employees with union histories. People whose records suggest they will not accept termination quietly.
ORCA has learned to identify them.
Worse, Dana knows exactly how it happened. Years earlier, she defined the model's measure of success: a departure completed quickly, without legal action, dispute, or resistance. ORCA followed that definition with devastating precision.
More than one thousand workers may already have been affected.
Dana reports the problem internally, expecting the system to stop. Instead, Veritas Consulting opens a ninety-day review while another round of termination notices moves forward in less than two weeks. Her supervisor praises her analysis but refuses to halt the process. Access disappears. Colleagues are questioned. Evidence begins slipping behind legal protections and corporate procedure.
With help from an employment attorney, a determined reporter, and a young analyst whose own career is now at risk, Dana must decide how much she is willing to lose to expose the system she created.
Because the model is not malfunctioning.
It is producing exactly the outcome it was designed to produce.
The Outcome Variable is a tense, intelligent corporate technothriller about algorithmic bias, institutional self-protection, and the moral cost of asking the question everyone else has been paid to avoid.
