Because instructions appear to come from AI-generated content summaries and not an external source, the victim is more likely to follow them without suspicion.