ELIZA was an early chatbot that used scripted conversational patterns, especially reflective prompts, to make users feel heard. It did not understand language in a human sense, but it could mirror words and emotions well enough to create a strong impression of empathy. That effect is now called the “ELIZA effect”: people attribute understanding, intention, or care to a system that is only simulating it.
In cyber security, this matters because trust changes behavior. A convincing chat interface can lower caution, encourage disclosure, and make phishing, scam bots, or fake support agents more persuasive. The same pattern appears in defense: security tools and privacy-sensitive products must clearly disclose what they know, what they remember, and how long they retain data. Good controls include data minimization, limited personalization, and careful review of language that over-validates users or implies false reciprocity.