It’s just a keyword-based filter at this point. Images/video would require AI interpreted scanning.