Type I error = False Positive (raising alarm when there's no threat).
Type II error = False Negative (missing a real threat — the more dangerous one).
Security parallels statistics: Type I (false positive) = over-alert. Type II (false negative) = under-alert. In biometrics: Type I = FAR (false acceptance, unauthorized user accepted). Type II = FRR (false rejection, authorized user rejected). Adjusting sensitivity trades one for the other.