A hybrid approach combining rule-based models with AI may be more effective in reducing false positives and identifying suspicious behavior.