This study developed and validated a new questionnaire, the LLM-D12, to measure human dependence and addiction potential for large-scale language models (LLMs). To overcome the conceptual limitations of existing scales that apply behavioral addiction symptoms to LLM use situations, the authors developed a 12-item questionnaire based on their previous theoretical research. Data collected from 526 UK participants were used for exploratory and confirmatory factor analyses, which identified a two-factor structure: instrumental dependence (6 items) and relational dependence (6 items). Instrumental dependence measures the degree to which individuals rely on LLMs for decision-making and cognitive tasks, while relational dependence indicates the tendency to perceive LLMs as socially meaningful, perceptive, and companion-like. The LLM-D12 scale demonstrated excellent internal consistency and clear discriminant validity, and external validation confirmed the distinction between the two subscales. We interpreted the psychometric properties and structure of the LLM-D12 scale considering that although LLM dependence does not necessarily indicate dysfunction, it may reflect a level of dependence that may be problematic in certain situations.
