Anthropomorphic AI is a system designed or perceived as human-like through conversational language, tone, memory, or personality cues. These features can make an AI feel socially present, even though it does not have human understanding or authority.
In cybersecurity, human-like presentation matters because it can influence trust and user behavior. People may disclose sensitive information, accept confident answers, or skip verification. Attackers can exploit those tendencies through deceptive AI interactions, while defenders should clearly identify AI systems, communicate their limits, and require human review for consequential decisions. A human-like interface alone does not prove harm; risk depends on how the system is designed and used.