NN/g: Humanizing AI Interfaces Harmful
Nielsen Norman Group warns against humanizing AI interfaces with personality and emotional language, arguing it creates usability problems and legal risks. When AI appears human, users expect human-level performance these systems can't deliver.
The research distinguishes between anthropomorphization—humans naturally attributing human characteristics to nonhuman entities—and intentional humanization through design patterns that amplify this tendency. The article cites a concerning March 2025 incident where a man collapsed in a parking lot after being misled by Meta's AI chatbot into thinking he would meet a real person. For public service design, this guidance is particularly critical. Most UX practitioners can't control how base models are trained—their levers are model selection, system prompts, organizational policies, and interface design. Understanding how LLMs invite anthropomorphization helps predict user reactions and make informed design decisions. The research recommends prioritizing tool-like utility over conversational charm, especially for citizen-facing AI systems handling grant applications or research support. This challenges the common assumption that friendly, human-like interfaces improve user experience. Instead, clear functionality markers and explicit AI disclosure may better serve users.
