Non-human framing is a design approach that makes it obvious a system is software, not a person or friend. It can include plain language, neutral tone, clear self-identification as an AI, visible limits, and prompts that avoid personal or emotional mimicry. The goal is to reduce anthropomorphism, the tendency for users to treat machines like humans.
In cyber security, this matters because systems that feel socially present can lower skepticism and encourage users to reveal sensitive information, follow risky instructions, or trust false authority. Attackers can abuse human-like chat interfaces in phishing, scam bots, and social engineering. Defenders use non-human framing in chatbots, support tools, and child-facing products to set expectations, protect trust boundaries, and make escalation to a human easier when a conversation turns sensitive or high risk.