Meridian's ramp-safety team deploys a computer-vision system that watches live camera feeds and alerts a supervisor whenever a ground-crew member enters a marked danger zone while an aircraft's engines are running. Which of the seven patterns of AI does this best represent?

CPMAI Task 2 requires developing pattern and anomaly detection capabilities and correctly matching them to business problems, and a ramp-safety zone-violation alert is a canonical example: the system is trained on what 'normal' safe positioning looks like and flags the anomaly (a person in a danger zone during engine operation) in real time. This is recognition/anomaly detection, not a different pattern. The 'conversational pattern' distractor confuses the downstream notification channel (an alert message a supervisor reads) with the core AI task, which is visual anomaly detection, not dialogue — the same mistake as pairing a detection system with a chat interface elsewhere in this domain. The 'hyper-personalization' distractor invents individualized preference-tuning that has no place in a safety-alert system, where the threshold should be a consistent safety standard, not a per-employee preference — personalizing safety thresholds would itself be a questionable and inconsistent safety practice. The 'predictive analytics, predicts the exact minute' distractor overstates the system's actual function; it detects a present-moment safety-zone violation as it occurs, it does not forecast a future accident time. Correct pattern identification also drives the right model type (vision-based anomaly detection) and the right response workflow (immediate alert, not a forecast report).