A Split Outcome: Two Engineers, One Decision, Very Different Judgments
A real-world example of Moral luck in action
Context
A mid-sized self-driving car company releases a software update intended to improve urban pedestrian detection. The engineering team used the same decision framework and testing regimen for all vehicles, but real-world edge cases are inherently rare and hard to reproduce in simulation.
Situation
Two senior engineers (A and B) each approved the same risk/benefit tradeoff for different vehicle batches. Both relied on identical sensor fusion thresholds and the same validation data; both documented the reasoning and logged the tests. Six weeks after deployment, a vehicle from Engineer A's batch was involved in a pedestrian fatality during an unusual lighting condition. No similar incidents occurred in Engineer B's fleet.
The bias in action
Public and internal reactions focused heavily on the tragic outcome, equating the engineer associated with the batch to moral culpability even though the documented intent and process matched that of the other engineer. Managers and the press treated the two engineers as morally different: Engineer A was criticized, demoted, and socially shunned; Engineer B was praised for 'good luck' despite having made the same choices. Investigations emphasized the outcome instead of rigorously comparing the decision processes, and performance reviews and HR actions reflected that skewed moral assessment.
Outcome
The company fired Engineer A, initiated a public apology, and settled a lawsuit with the victim's family. Engineer B received a token award and continued in a leadership role. Internally, trust eroded: engineers began making more conservative decisions to avoid being the person associated with a negative outcome, slowing product delivery and innovation.




