A machine learning engineer explains that their malware classification model uses a 'confusion matrix' for evaluation. In a binary malware/benign classifier, which cell of the confusion matrix represents the most dangerous error type in a production endpoint security context?

• A.True Negative: benign file correctly identified as benign, because model files are digitally signed by default in every major registry
• B.True Positive: malware flagged as malware, since a confusion matrix is only defined for perfectly balanced classes
• C.False Negative: malware file misclassified as benign - the model fails to detect a real threat
• D.False Positive: benign file flagged as malware - causes analyst workload but no security gap