A new arXiv preprint asks a pointed question: does the presence of algorithmic review change the decisions a system is meant to correct? The researchers use professional tennis as a natural experiment, where Hawk-Eye technology can overrule umpire line calls. By comparing umpire decisions in situations with and without review, they aim to isolate how the threat of correction shapes behavior.
The design is notable because it shifts attention from post-hoc error correction to the upstream effects of oversight. If umpires anticipate algorithmic scrutiny, they may adjust their calls in advance—becoming more cautious, more accurate, or perhaps biased in other ways. The abstract does not report specific results, so the magnitude or direction of any effect remains unclear from the summary alone.
Still, the research points to a broader lesson for human-AI systems: algorithms do not just audit human judgment; they can also reshape it. The paper is currently a preprint, meaning the findings should be treated as preliminary until peer review and replication.